Abstracts

A metric of color space for science, industry and art

Karl R. Gegenfurtner, Andrea van Doorn, Doris Braun, Jan Koenderink

Justus Liebig University Giessen

It is well established that color space is three-dimensional, but much less is known about its perceptual granularity. Vision scientists have long mapped small regions of it by measuring just-noticeable differences (JNDs) between pairs of colors, leading to estimates of millions of discriminable colors under ideal conditions. Artists and designers, by contrast, have relied on color atlases constructed from qualitative steps—palettes of only a few hundred colors that feel equally distinct. We sought to bridge these two approaches by empirically determining a full metric field across the RGB cube, the restricted space that defines most digital imagery. Instead of measuring JNDs, eight observers adjusted the color of a test patch until it appeared notably qualitatively different (NQD) from a fixed reference color. Thirty-five reference colors, arranged in a body-centered cubic lattice, sampled the cube evenly. For each reference, NQDs were determined along seven directions connecting it to its nearest neighbors. The resulting 35 local ellipsoids describe the local geometry of perceptual color differences and together characterize the metric structure of RGB color space. There was remarkable agreement between observers in the shape and orientation of the ellipsoids. Overall sizes differed due to individual criteria, but simple scaling brought them into close correspondence. By visualizing this empirical color-space metric, we reveal both its geometry and its limits: the space of experienced colors is not infinite but surprisingly coarse, containing roughly 250 distinct colors for the average observer. Our results shows that direct measurement of the metric of color space is possible, overcoming the “curse of dimensionality”. The resulting structure provides a common map of color space for vision science, design and artistic practice.

Being Malevich II: Color, Balance, Dynamic and Liking in Suprematist Compositions

Doris Braun, Vanessa Kremer, Katja Dörschner

Giessen University

What makes a composition compelling? We examined how color and compositional intent modified compositions and aesthetic judgments of stability, dynamics and attractiveness in Suprematist artworks by combining hands-on construction tasks with perceptual ratings. Stimuli were derived from ten Suprematist artworks. For each, we prepared two sets of paper stimuli: one preserving the original palette and background, and an achromatic one with black shapes and white backgrounds. Twenty participants created three versions per artwork: a stable, dynamic, and interesting one (yielding 30 compositions per person) by arranging first paper cutouts and then transferring them to the computer. Then participants rated their own compositions and the originals on balance, stability, dynamism, liking, and interest. We characterized the layouts with quantitative descriptors of symmetry, mass distribution, angular deviation from cardinal axes and spatial dispersion. Stable compositions were more symmetrical with a lower center of mass, often “stacked” toward the bottom. Dynamic compositions showed larger angular deviations and wider spatial spread; these deviations were slightly greater in color than in black-and-white. “Interesting” compositions also exhibited greater diversity of orientations and positions, but we found no significant difference for color. In the ratings, “interesting” compositions were liked most and judged most balanced. Across tasks, color increased liking and perceived interestingness. Color also enhanced the perceived balance of stable arrangements, amplified perceived dynamism in the originals and in dynamic compositions, but did not change perceived stability. Individual participants differed in their preferences for stability versus dynamism, and these preferences shifted with color. Together, the results suggest that perceived interestingness is supported by structural variety in orientation and placement and color primarily acts as an affective amplifier and selectively modulates balance and dynamism. Our findings connect principles of perceptual organization of artworks and demonstrate the value of participatory construction paradigms and individual preferences.

Blue-Induced Temporal Dilation in Art Viewing

Cehao Yu, Müge Cavdan

Aarhus Univerisity

Illumination shifts the apparent chromaticity of artworks. Here, we tested whether global white-point shifts bias perceived viewing duration of paintings, and how such shifts relate to affect and aesthetic appraisal. Five Van Gogh paintings were digitally reproduced and rendered in five chromatic conditions: Neutral plus white-point shifts toward red, green, blue, and yellow. Shifts were matched in ΔE00 with D65 as the reference white, and mean brightness was held constant. 50 observers completed a temporal bisection task (comparison intervals 400–700 ms), from which we estimated the point of subjective equality (PSE) and just-noticeable difference (JND) and Weber fraction via psychometric fits. Afterwards, each stimulus was rated on Aesthetic quality, Artistic merit, Discomfort, and Pleasantness. Perceived durations depended on chromatic conditions. Blue-shifted condition showed significantly lower PSEs than Neutral and Green (judged longer; ps < .05). JNDs/Weber fractions did not differ significantly across chromatic conditions (precision unchanged). Aesthetic quality correlated positively with Artistic merit, and higher Pleasantness was associated with lower Discomfort. Across stimuli, higher PSEs were associated with greater Discomfort (r = .47, p < .05) and lower Pleasantness (r = −.47, p < .05). Mean Aesthetic ratings did not differ significantly between chromatic conditions (repeated-measures ANOVA and post-hoc tests, all ps > .05). Blue white-point shifts induce a time-dilation bias for paintings without impairing temporal sensitivity, consistent with a perceptual bias rather than altered discrimination. The association between longer perceived time and higher pleasantness/lower discomfort indicates that affect may contribute to chromatic timing biases at sub-second scales. Given the brief exposures, these effects are more plausibly linked to fast visual pathways than to slower melanopsin-mediated mechanisms. Illumination-induced chromatic shifts can thus shape time perception during art viewing, with implications for museum and well-being–oriented lighting design.

Cloth between the wrinkles: exploring information propagation in surface perception

Celine Aubuchon, Roland W Fleming

Justus Liebig University Giessen

One of the great achievements in figure sculpture is the compelling depiction of cloth. Blocks of solid marble are transformed through meticulous shaping into delicate folds of linen. Cloth itself has no recognizable form; its shape—the way it folds, crumples, and drapes— entirely depends on interactions with the physical world. Accordingly, aside from wrinkles and folds, a cloth may be invisible on a sculpture’s surface. But between wrinkles, the figure does not appear naked. This references a remarkable ability of our perception: to make inferences given incredibly sparse information.

In our experiments, we tested the limits of this ability by measuring where cloth perception begins and ends on surfaces with extremely minimal cues to cloth. To do this, we rendered images of virtual sculptures, each shaped as an unrecognizable ‘blob’, which either included sparse cloth-like wrinkles (cloth condition) or did not (clothless condition). This was such that there were only a handful of pixels that were different between the cloth or clothless version of a sculpture. Participants reported whether or not they perceived cloth at each location in the image.

Unsurprisingly, in the clothless condition, participants did not perceive cloth anywhere on the surface. But the sparse wrinkles in the cloth condition had far-reaching influence: cloth appeared to cover the sculpture as if it were wrapped. This is especially striking considering that most of the areas that were perceived as cloth-covered were pixel-for-pixel identical to the perceptually bare clothless version. This hints that the perception of cloth—and indeed many other surface properties—may be driven by a propagation of information across the image where sparse image cues are interpolated to fill in a surface.

Controlled Instability: Graphic Design as Embodied Interaction in Real-Time Systems

Marvin de Jong

Independent Designer

This work positions graphic design within three-dimensional, interactive space. Camera position, focal length, lighting, and surface materiality act as variables alongside letterforms, creating environments where depth and scale become ambiguous. The viewer’s position and movement shape not only what is seen but how spatial relationships are experienced.

Through MIDI controllers, gesture recognition, and environmental sensors, these systems translate physical input into dynamic visual fields. Form and perception operate as parameter-based relationships rather than fixed compositions.

Drawing on Merleau-Ponty’s embodied phenomenology and Lee Ufan’s concept of relatum, the projects explore form as encounter, where design is positioned as a felt dialogue between system, viewer, and space. These real-time systems reveal how emerging design tools can lead to further explorations of our relationship to form and perception, creating environments where design, space, and the body converge as unified fields of interaction.

Cultural Differences in Gaze Patterns over Paintings and Photographs

Chao-Shan Hsu, Xingyu Long, Raphael Rosenberg

University of Vienna

How culture shapes the uptake of pictorial form has long been argued in art history, anthropology, and psychology, yet few studies combine spatially mapping with per-second gaze dynamics across visual media. We report a cross-cultural eye-tracking study comparing Austrian (AT) and Japanese (JP) adults (Brinkmann et al., 2023). Participants freely viewed Japanese and European paintings and photographs for 12 seconds. Each image was partitioned into 4×4, 8×8, and 16×16 grid cells (treated as AOIs); fixations were assigned to grid cells and saccades tallied as grid-cell transitions. For paintings, figure and ground were manually segmented to compute gaze share. We quantified (a) per-AOI fixation share (time-normalised); (b) within-group similarity of fixation and saccade distributions; and (c) a scanpath-mode K-coefficient derived from fixation duration and saccade amplitude (lower K = ambient/exploratory; higher K = focal/inspection). Three findings stand out. (1) JP showed higher within-group similarity than AT for both fixations and saccades across grid resolutions and stimulus blocks. (2) Differences were spatially specific rather than uniform; a subset of AOIs showed AT–JP gaps >10 percentage points, and in figure–ground paintings JP devoted higher fixation density to figure/foreground AOIs than AT. (3) Both groups shifted from ambient to focal around 4–5 s, but JP’s shift was weaker and more stable; JP also differentiated paintings and photographs more strongly (shorter saccades for paintings, longer for photographs, with smaller effects in fixation durations), whereas AT were comparatively medium-invariant. These results reveal a position-specific, time- and medium-dependent cultural signature in picture viewing, suggesting strategy-level differences in how formal structure is sampled. Future work should test mechanisms such as expectation and task framing, and broaden to additional cultural groups to assess generality and boundary conditions.

Depicting human figures in line: comparing expert vs. beginner decision making in tracing a figure from photo reference

Mark Sypesteyn, Sylvia Pont, Maarten Wijntjes

Delft University of Technology

For many design students the practice of drawing remains a relevant part of their education. With the field of design expanding from just the physical product design into the realms of social design, human centred design and many more (Corremans & Mulder-Nijkamp, 2019; Rodriguez & Peralta, 2014), depicting the figure is becoming increasingly relevant to the design student and, as a consequence, drawing the figure is as well. Yet within the practice of drawing, the subject of the human figure often proves to be exceedingly difficult for learners to practice and master. Previous works investigating the differences between expert and layman drawings have had limitations. Most works do not focus on the human figure as reference (Bainbridge et al., 2019; Brady et al., 2008) and show several limitations. For example, differences in observation behaviour (B. Howell et al., 2023; B. F. Howell et al., 2024), visual memory or reference discrepancies were not compensated for. Besides, either the subject to draw or the elements with which to draw were not isolated (Cohen & Bennett, 1997). In this study we isolate the line as drawing element and ask the question whether we can distinguish beginners and advanced traced figures – and which lines bear the most value when depicting the figure. An experiment was conducted in which 20 beginner drawings are compared to 20 advanced drawings. The task for all 40 participants was to produce a series of line drawings, tracing over the same six reference photographs. Besides obvious commonalities, the first three sets of drawings show a wide variety in line drawings and decisions – between layman v. expert groups and within the groups.

Diego Velázquez: Image within an Image

Doroteja Ivanec, Ana Mišković

University of Zadar

The issue of idea and/or motif of the “image within the image” is ubiquitous in modern and contemporary art represented mostly as metapainting. Interestingly, even Baroque painters sometimes dealt with the same issues, probably unintentional.
The aim of paper is the recurring idea and/or motif of the “image within the image” in paintings of Diego Velázquez, a significant Spanish Baroque painter and influential European visual theorist/thinker. The study examines how Velázquez employs visual and spatial structures to pose questions about vision, the act of seeing, the painting process, and the viewer’s position in addition to depict narrative content through close analyses of “Christ in the House of Martha and Mary”, “The Rokeby Venus”, “Las Hilanderas”, and “Las Meninas”. The artworks turn into a place where painter and viewer, reality and representation, meet. Authors consider how Velázquez’s manipulation of light, space, materiality and gesture enhances a theory of depiction, representation and perception. We address the tension between the illusory and the tangible, the earthly and the transcendent - an idea that bridges readings of Baroque drama e.g., P. Calderón de la Barca, and the complex spatial layering in Velázquez’s paintings. The paper pulls from a theoretical framework that encompasses the works of intellectuals like Michel Foucault, Svetlana Alpers, and Sam Rose. The theoretical frame argues that Velázquez’s “image within an image” becomes a reflection on the processes of creating, viewing, and interpreting images that function beyond a formal technique. Velázquez does more than paint the world; he poses questions about how we see it and how it is created through the idea and/or motif of the “image within an image”. By doing this, his work transcends its historical setting and continues to be important as a type of visual philosophy.

Enlivening Beautiful Forms. New Insights into Empathy and the Aesthetics of the Visual Arts

David Romand

Aix Marseille University

This presentation aims to revisit the role of empathy in visual aesthetics, that is, our capacity to apprehend visual artworks or parts of them as “enlivened” when contemplating them. First, I propose a critical survey of old and new approaches to empathy and visual aesthetics (Lipps, 1903, 1906; Freedberg and Gallese, 2007). Here I stress the need to elaborate a well-defined typology of the empathetic processes involved in the aesthetic apprehension of visual artworks by specifying what kinds of visual forms are likely to be empathized with in them. Second, I address the question of the definition of empathy, which I identify with our power to subjectivize the constituent elements of the external world, whether perceived, imagined, or recollected. I insist that, when contemplating visual artworks, what we empathize with are all perceptual contents that we experience together with a feeling of animacy. Third, I show that, in the visual arts, there are three chief categories of perceptual contents we empathize with, three kinds of visual forms that can be arranged according to their decreasing degree of depictability: (a) all or parts of human and nonhuman figures; (b) perceptual contents that, without being identifiable as well-defined living things, characteristically evoke the presence of biological entities or events; (c) sensory spaces with blurred boundaries that characteristically elicit an “atmospheric” feeling of subjectivity. As I highlight, each of these three visual forms is the object of a specific kind of empathy, namely, (a’) “figural empathy,” (b’) “biomorphic empathy,” and (c’) “perceptual mood empathy,” of which I analyze the nature, as well as the respective place in visual aesthetics. Fourth, I conclude that empathy is a substantial component of visual aesthetics, whose significance depends on the formal organization specific to each artistic domain and to each artwork in particular.

Eye Tracking Perception of Traditional, Impressionist, and Cubist Still Life Paintings at the Manchester Art Gallery

Christopher Linden, Aleksandra (Sasha) Igdalova, Rebecca Chamberlain, Johan Wagemans

KU Leuven

Across stylistic evolutions, from the Dutch Golden Age to impressionism to contemporary abstract art, still life has remained a mainstay genre in western painting. The symbolism rife in these works may not always be clear to gallery visitors, especially for less traditional styles. We partnered with the Manchester Art Gallery (MAG) to explore how different content and presentation formats of guides impact viewers’ perception of still life paintings across different artistic styles. Our 216 participants viewed three still life paintings for four minutes each, while their eye-movements were recorded using mobile eye tracking (MET) glasses. The third-viewed artwork was always self-selected from a set of 13 traditional still life paintings in the MAG’s Dutch gallery. The initial two artworks were counterbalanced in viewing order between an impressionist still life, by Edward Hartley Mooney, and a cubist still life, by Ben Nicholson, both viewed in MAG’s Room To Breathe gallery. The first-viewed artwork acted as a baseline, presented without information. The second was presented alongside guiding information, with a 2x2 manipulation of guide content (art historical vs. visual thinking) and guide presentation (audio vs. live). The third (self-selected) artwork acted as a test condition, viewed without information to assess whether the guide had forward transfer effects on visual exploration and art appreciation. After viewing each artwork, participants answered a series of questions about their aesthetic and emotional responses to that painting. The guiding information had a positive effect on participants aesthetic and emotional responses to the artwork, which also seemed to carry forward to the third-viewed artwork. Scan-path similarity analyses will be used to examine the common viewing patterns within and between the individual artworks, as well as the extent to which viewing patterns aligned across participants during and following guided information.

From Formal Features to Discursive Vision: Distinguishing and Articulating Human and Machine Perception of Art

Darío Negueruela del Castillo, Ana Zapata, Iacopo Neri, Andrea Alfarano, Eduardo Trabattoni

University of Zurich

Multimodal vision-language models can describe images fluently, but they struggle when faced with artworks. This paper explores this the gap isn’t just about accuracy, but reflects a fundamental difference in how humans and machines perceive visual form. When we look at paintings, we don’t simply catalog objects and attributes. We read spatial tensions, trace how light sculpts volume, notice where a contour buckles or dissolves. We bring aesthetic expectations shaped by centuries of looking, making, and theorizing about images. Machines, by contrast, extract patterns optimized for general visual tasks, which succeeds with photographs but fails when confronted with the deliberate distortions, symbolic densities, and material intensities of artistic practice. Using VQArt-Bench, a semantically rich benchmark for art and cultural heritage, we show how current models exhibit a peculiar asymmetry: they generate plausible art-historical discourse without adequately grounding it in what’s actually visible. They miss basic formal relationshipshow space is constructed, how forms interact, how stylistic choices create meaning. This reveals that visual arts are not just harder examples within existing visual categories, but they operate according to different perceptual logics. To address this, we draw selectively on aesthetic theory, particularly baroque concepts of deformation and excess, alongside frameworks emphasizing the materiality and historical situatedness of visual form. These perspectives help articulate that machines miss the capacity to recognize artworks as dynamic events rather than static inventories of features.We then propose fine-tuning a lightweight multimodal model specifically for artistic visual reasoning. Our approach combines form-aware contrastive training with multi-agent generation of grounded explanations, evaluated across new axes including formal differentiation, contextual sensitivity, and interpretative stability. Ultimately, this work argues that meaningful comparison between human and machine perception in art requires moving beyond accuracy scores toward understanding how visual form becomes significant.

From Seeing to Unseeing: Expanding the Vision and Depiction Framework Through Making

Catelijne van Middelkoop

TU Delft & University of the Arts The Hague

This presentation explores how the Vision and Depiction framework (Wijntjes & Van Middelkoop, 2024) can be extended through making as a didactic principle (Van Middelkoop, 2025) to deepen students’ perceptual and conceptual engagement with images. In our curriculum, the framework initially supports students in learning how to see: by analysing curated images, students investigate formal elements, perceptual cues, and the visual strategies that shape meaning. Formal analysis, combined with information gathered through reverse-image searches, provides a rich starting point that fuels their curiosity. Students then translate these insights into a poster presentation based on a research question they formulate and operationalise through small-scale experiments. This process brings them into close contact with the visual and conceptual mechanisms at play. However, while students can articulate what an image does, they often struggle to create visualisations that do the same—guiding a viewer’s eye with intentionality. This struggle reveals a missing dimension: experience in making. To address this, we introduce an additional assignment in which students reenact the making process of their selected image. Through recreating the work, using materials and techniques ranging from paper crafts to generative AI, students learn from divergence. They produce many versions, compare them, and analyse how each variation shifts perception. Individual images from the curated set thereby transform from isolated artefacts into catalysts for material exploration and embodied understanding. This experiment includes both fine-art students and those uncomfortable with making. For all of them, the shift is significant: the aim is not just to help students see, but to bring them to a point where they cannot unsee. Understanding the process of making opens a space where vision, depiction, and imagination become intertwined. In this expanded framework, making becomes essential for learning to see, and for envisioning what might yet be made.

Gold, Reflection and Glare: The Paintings of Lovro Artuković and The Logic of Paradox in Mimesis

Nikola Zmijarević

University of Zadar

This paper examines how the painting cycle “Zlato, odraz i odsjaj” (eng. Gold, reflection and glare) by contemporary Croatian painter Lovro Artuković, stages a dialogue between formal optics (light, texture, specular reflection) and painterly material (visible brushstrokes), thereby addressing the interplay of vision and depiction across motif and medium, content and form. The paper analyses how Artuković renders specularity—edge glare, flare, and reflective distortions—through pictorial means that are themselves non-specular, so that the depicted reflection and the paint film diverge. The works thus produce what Lacoue-Labarthe calls the logic of paradox, which is the very logic of mimesis: the more something resembles another thing, the more apparent is their difference (L’imitation des modernes, 1985). In Artuković, this mismatch compels the viewer to complete an optical fiction that the surface itself denies.

While situating Artuković within a historical tradition of mirror depiction in Western painting, from Van Eyck and Parmigianino, Caravaggio and Velázquez, to Manet and Bonnard, the paper shows how the painter updates the mirror’s semiotic charge (self/other, aura/commodity, reality/illusion, identity/difference) for a postmodern, post-photographic context. It argues that these paintings demonstrate the impossibility of mimetic reproduction at the very point of its greatest plausibility, and render seeing itself – its instabilities, subjectivity, and temporality – as motif, while using the medium to materialize that experience. Methodologically, the paper combines formal analysis (colour contrasts, highlight structure, paint application, faceted surface fragmentation) with visual semiotics (mirror as sign) and a selective optics vocabulary (specular vs diffuse reflection). The result is a case for how contemporary painting can bridge science (philosophy) and art: by turning optical phenomena into symbolic forms (Cassirer, Panovsky), so that the physics of light becomes legible as cultural meaning.

How Rhythms Shape Meanings in Cinema: A Videographic Gestalt Approach

Maarten Coëgnarts

LUCA School of Arts / University of Antwerp / FilmEU

This paper extends Rudolf Arnheim’s Gestalt approach—originally developed for the fixed arts—to the domain of moving images. In Art and Visual Perception ([1954] 1974) and Visual Thinking (1969), Arnheim employed schematic diagrams to uncover the underlying “structural skeletons,” or Gestalts, that serve as carriers of meaning intended by the artist. He challenged the dualistic separation between perception and thought, arguing that an artwork’s formal organisation—its shapes, lines, colours, and composition—is inseparable from its meaning. For Arnheim, form itself is expressive: even abstract configurations can evoke emotion and significance purely through perceptual structure (see also Winner 2019: 64). “In great works of art,” he writes, “the deepest significance is transmitted to the eye with powerful directness by the perceptual characteristics of the compositional pattern” (Arnheim [1954] 1974: 458).

While Arnheim illustrated his analyses with diagrams and sketches that reveal hidden structural relations, these tools were confined to static forms such as painting or sculpture. Cinema, however, adds the dimension of time—its meanings unfold through movement, rhythm, and transformation. To address this temporal challenge, the paper proposes a schematic and animated approach to film analysis using the graphic tools of the medium of the video essay (Coëgnarts 2023, 2025; Kiss and Coëgnarts 2025). Through keyframe animation and overlays, abstract shapes and vectors can trace rhythmic patterns of motion and containment across film scenes. Drawing upon the author’s extensive research in this field (e.g., Coëgnarts 2025; Coëgnarts and Slugan 2022), the paper will present a series of case studies that demonstrate how such animated schematics reveal rhythm as a fundamental Gestalt principle in cinematic meaning-making.

How smear frames improve fast motion legibility

Jean Basset, Pierre Bénard, Pascal Barla

Université de Bordeaux

In 2D and 3D animations, artists use different stylization techniques to expressively convey motion. Among these techniques, smear frames aim to exaggerate motion by emphasizing the trajectory of animated objects through deformations. One important goal of smear frames is to improve fast motion legibility by guiding the gaze of observers during animation. In this work we explore conditions where smear frames effectively help guide the gaze.

We created simple 3D animations of a ball moving left to right on a screen, with varying trajectories, velocities and acceleration patterns. For each animated sequence, we also created two additional versions with smear frames of increasing intensity. Specifically, we used elongated in-betweens, where the ball is stretched to cover its position in the one or two previous and next frames of the animation. We ran a psycho-physics experiment in which participants were instructed to follow with their gaze the center of the ball during these animations. We measured the position of their gaze on screen with an eye-tracker.

We use the data collected during this experiment to study the ``error’’ of the participants, computed as the distance between the gaze position and the actual position of the ball at each timestamp. We observed that at high velocities, animations with higher smear levels have lower mean errors. For high accelerations, higher smear levels have lower error variance. These first results confirm that smear frames help guiding the gaze of observers in fast and rapidly accelerating motions. In further explorations, we plan to establish speed and acceleration thresholds above which smearing becomes relevant, and explore smooth pursuit abilities as they are linked to a good tracking performance.

Iconography of the Digital Self: From Beauty Filters to AI Portraits

Piera Riccio, Miriam Doh, Nanne van Noord, Nuria Oliver

University of Amsterdam

Recent advances in generative AI portraiture show a shift towards resembling traditional beautification practices, resulting in AI-generated portraits that increasingly look like enhanced selfies. This convergence suggests that generative AI portraits and beauty filters exist along the same continuum of digital self-presentation, where the lines between “filter” and “portrait,” and between real-time augmentation and generative synthesis, are progressively harder to distinguish.

We present a comparative study of digital beautification between beauty filters and AI portraits, examining how each adopts the conventions of portraiture. Beauty filters, a pervasive phenomenon on online platforms, rely on facial recognition and augmented reality to extend a long history of pictorial idealization. AI portraits, in parallel, use diffusion-based generative AI, inheriting historical aesthetic logics while introducing novel forms of representation.

The human face, and particularly the skin as its material surface, is central to this analysis, carrying iconographic and sociological significance. Beauty filters intervene at this level of materiality (modifying skin texture, radiance, and uniformity, while also directly “adjusting” facial features) to produce a digitally optimized performance of the self. In their early stages, instead, AI portrait tools often produced images that were either limited in realism due to technical constraints or intentionally stylized for aesthetic effect. Many applications allowed users to imagine themselves as digitally mediated “icons” that conveyed aesthetic forms of social status. Over time, AI portraiture has shifted closer to traditional beautification, with generated images increasingly resembling realistic selfies and integrating AI’s generative power into everyday self-presentation.

We investigate the continuum between beauty filters and AI portraits by situating these practices within the history of portraiture, and the broader context of algorithmic fairness and representational biases. We argue that digital beautification represents a domain where social aspiration and iconography intersect, redefining the culture behind self-presentation.

Illusory Colours in monocular rivalry, new paintings.

Leone Burridge

Independent artist

At Visio and Depiction 2024 i presented paintings made with only 2 colours and grey, but which gave percepts of additional colours alternating in monocular rivalry. The paintings were made with fine lines of the 2 colours and grey to create additive colour mixing. With the alternating percepts of illusory colours, it seemed that there was an alternation between assimilation of colour and simultaneous contrast of colour.

In this 2026 Vision and Depiction presentation the paintings are made with physical mixtures of the 2 colours and grey in subtractive colour mixing. There are no fine lines of colour which would give rise to assimilation of colour but the alternating illusory colours can still be seen. This raises further questions as to what may be the underlying perceptual processes underlying this phenomenon.

In the artist’s eye: manuals on painting techniques, written by artists for artists as a tool for understanding the creative process.

Ilaria Paolucci

Sapienza Università di Roma

Between the late Nineteenth and early Twentieth centuries, a growing number of manuals on painting techniques were published throughout Europe. This genre of publishing has the rare characteristic of examining painted works of art and the repertoire of technical processes used to create them, from the artists’ point of view. In fact, it was often the artists themselves who wrote the manuals, with the stated aim of providing practical guidance to new generations. One of the reasons for the growing demand for these books was the artists’ dissatisfaction with academic teaching and the increasingly widespread practice of painting en plein air. Preferred subjects were the phenomena of light as they appeared, in a struggle between conventional pictorial techniques and the mimetic rendering of visual perceptions. This paper examines a selection of theses, taken from manuals on painting techniques published in Italy in the context of Divisionism (1891-1920), which focus on describing the process that takes place in the artist’s mind when, drawing from life, he or she must represent what is observed. Controversial teaching practices will be highlighted, along with the exercises suggested for achieving the desired pictorial effects. The selected texts come in particular from the treatise “La science de la peinture” by J.G.Vibert (1891, translated into Italian by Previati), which was an undisputed practical guide for the movement’s theorist, V.Grubicy de Dragon, from articles and letters by him, and from Previati’s treatises, the first of which was “La tecnica della pittura” (1905), which became reference texts for Italian painters. Finally, the legacy of these treatises will be highlighted in L.A. Rosa’s “La tecnica della pittura” (1937) which was completed after the First World War, and can be considered the last repository of this genealogy of writings.

Inside-out dynamics of art perception through a multi-method lens

Eleftheria Pistolas

University of Leuven

This work is centered on the idea that the experienced world is not merely perceived but actively constructed, highlighting the dynamic role of the perceiver in vision and aesthetic appreciation. In this talk, I will explore how perceived form interacts with the intentions of the artist, the physical medium, and the embodied engagement of the viewer. Across two research lines, we investigated the dynamic interplay of stimulus, context, and perceiver in perceptual and aesthetic experiences. The first line of research is focused on perceiving from within and examines immersive Ganzfeld environments as paradigms of inward-directed perception. When external structure and changes in visual information are minimized, observers generate their own perceptual world containing colors and shapes, altered depth, floating sensations, and shifts in states of consciousness. These experiments allow us to reflect on how a constrained medium, i.e., a uniform field of light, can give rise to rich perceptual experiences, revealing the perceiver’s constructive contribution. The second line of research is focused on inside-out dynamics in visuo-tactile engagement with sculptural art. Perceivers approach sculptures with particular exploratory aims, guided by what they seek to experience. Through manipulation of visual exploration and tactile interaction, this work shows how different modes of exploration draw attention to distinct aspects of form, such as visual potential versus material properties. Throughout these lines of research, we have combined neural (EEG) and behavioral (eye-tracking) measures with rating scales and qualitative reports for a more integrated understanding of the studied perceptual experiences. This talk will illustrate how the kinds of questions that lie at the crossroads of different disciplines can benefit from integrating multiple types of data, ranging from neural and behavioral to qualitative data for a more comprehensive and multi-layered understanding of how we perceive and appreciate aesthetic experiences.

Interrogating Design as a Cultural Metatool

Ryan Pescatore Frisk

Independent Researcher

This presentation examines material ideologies as biases inherent in semiotic networks within the research and development of everyday artifacts in an increasingly interconnected world. It begins with a review of the elements and principles of design, as fundamental patterns of structures, and their interactions, comprising visual and material representation, drawing from professional, pedagogical, and scientific disciplines. Building upon this knowledge, the discussion incorporates concepts from social semiotics, positioning design as a counterpoint to fixed and canonical compositional frameworks.

Material manifestation is a central component of human culture, where value in everyday artifacts is continually and incessantly made meaningful through social acts. As everyday life becomes increasingly permeated by shared sociotechnical systems, it also becomes more intertwined with shared visual, digital, and material culture. This interconnectedness empowers highly stratified human identities, practices of use, and subsequent social meaning. Design in today’s world, both in the processes that facilitate outcomes and the social practices where meaning is made, is distributed and highly intertextual, drawing on networks of potential across scales of time and space.

The study presents a collection of case studies and examples that interrogate ideas of design as a cultural metatool central to the conceptual, material, and constructed experiences that encompass various facets of human life, including the built environment, embodied experiences, temporal experiences, tactile experiences, pictorial experiences, auditory experiences, and across multi-sensory, transdisciplinary, multimodal, and transmodal entanglements. The study posits that the foundational competencies that humans employ to navigate, participate, and comprehend everyday practices establish a link between value derived from shared sociotechnical systems and ideas of design as a sort of cultural software. This connection facilitates the production of social affiliation and distinction, ranging from vernacular cosmopolitanism to the technological determinism of digital infrastructure and cultural non-spaces of generative AI.

Investigating the Effect of Child-Written Descriptions Using Eye-Tracking at the Groninger Museum

Océane Tilman, Selin Isik, Zsofia Pilz, Francesco Walker

Leiden University

Does descriptive context influence how both children and adults experience art? Previous research shows that painting descriptions can affect how adults and children view art (Walker et al., 2017), and that age-tailored descriptions can effectively influence children’s viewing behaviour (Walker et al., 2024). The Groninger Museum offers a unique approach to describing artworks, providing descriptions that articulate the artworks through the perspective of children. This study analyses three paintings and their corresponding child- and conventional descriptions at the Groninger Museum. It explores the effectiveness of child-written descriptions for both children and adults, compared to conventional descriptions. In the current study, participants listened to a description of a painting and then viewed the respective painting. This process was repeated for each painting, with the order of the phases counterbalanced. In one phase, participants listened to descriptions written for adults, and in the other, they listened to descriptions written by children. Mobile eye-tracking was used to measure gaze patterns across the entire painting and within predefined description-specific areas of interest (AOIs). Participants additionally responded to a short interview to assess recall and completed a questionnaire on aesthetic appreciation. We hypothesize that child-written descriptions will influence the viewing behaviour of both children and adults, who will spend more time viewing child-written description AOIs compared to conventional description AOIs. We also expect a greater recall of child-description AOIs among children. Similarly, we expect self-report measures of art appreciation to show that child-written descriptions evoke more positive emotions. The study aims to provide insight into how museums can incorporate child-written descriptions to enhance the individual experience of museum visitors across age-groups.

Latent Space, Through the eyes of the algorithm

Cathalijne Postma, Debbie van Berkel, Jojanneke Postma

studio KASBOEK

Latent Space is an interactive art installation that investigates how perception emerges in the interplay between human vision and artificial observation. When you see yourself in a mirror, you encounter only the surface. Latent Space gives the inner voice a chance to speak, revealing what is mirrored beneath rather than what simply appears above. The installation uses a mirrored interface that functions analogously to the human retina, a surface that catches your gaze and translates it into impulses that travel toward a deeper perceptual field. This in-between zone, the Latent Space, is where new connections arise. Here, AI agents communicate not as programmers but as observers, tracing how your eyes move across your mirrored image and interpreting this motion as a neurological pattern: rhythm, direction, focus, dispersal. Your way of looking becomes legible as a form of language, generating a digital reflection shaped not by appearance but by perceptual behaviour. In this reciprocal loop, the system mirrors you back in transformed form, foregrounding the invisible gestures of seeing that normally remain unnoticed and inviting viewers to reflect on how meaning emerges in the space between observer and observed, between the physical mirror and the computational processes that reinterpret it. By engaging participants in this hybrid perceptual loop, Latent Space challenges conventional boundaries between medium and motif, image and gaze, yourself and reflection. It positions AI not as a tool but as a co-observer, offering an expanded understanding of how vision can be depicted, interpreted, and mirrored back through the eyes of the algorithm.

Likin art; Light induced kinetics on art

Sylvia Pont, Katja Doerschner

TU Delft

Changing colors of illumination on colored artistic imagery can induce powerful illusions of a wide variety of perceptual qualities. We found dynamic illusions in which the (static) imagery seemed to glow, deform, wiggle, rotate, flow, change in depth, or material due to light variations. One type of rotation effect was tested using formal psychophysical methods [1]. The data could be well predicted by a quantitative coarse-grained (RGB channels) spectral model based on light-material interactions and effective luminance contrast. The other illusions were tested qualitatively, which confirmed the above dynamic qualities of the illusions [2]. The basic triggers for the kinetic illusions are contrast variations and even reversals in the effective proximal visual stimulus, caused by distal color combinations of print colors and illumination color that vary in their coarse grained spectral tuning. We will demonstrate a variety of such illusions and provide a setup in which participants of the Vision and Depiction meeting can try their own or other people’s art. We will also provide an “EXPLORA kit” with slides as lights and stimuli for them to be able to demonstrate the effects and test new images anywhere.

[1] Pont, S., & Doerschner, K. (2025). Changes in illumination color induce powerful illusory rotations. JOSA A, 42(5), B124–B132. https://doi.org/10.1364/JOSAA.545157

[2] Doerschner & Pont. To appear in a new light: varying illumination colors induce dynamic transformations of perceptual qualities in still image. Submitted.

Making Sense of Meaning-Making: Applying the Predictive Processing Framework to Art Photography

Tomas Vandecasteele

KULeuven/Luca School of Arts

This paper examines how meaning attribution and aesthetic value transcend the purely visible information in art photography, despite its strong resemblance to its referents. Using three case studies—Eadweard Muybridge’s Animal Locomotion (1879), Thomas Ruff’s Porträts (1981–1991/1998–2001) and Wolfgang Tillmans’ LED Flicker (2018)—I develop an argument that meaning does not coincide with indexical transparency or mimetic resemblance, but arises dynamically through inferential processes on the part of the viewer. Muybridge’s images dissect movement into sequences, thereby suggesting temporality and causality. What we “see” is not the horse as the eye sees it, but a temporarily segmented reconstruction of movement—phases that remain fundamentally invisible to the naked eye. Animal Locomotion exposes the epistemic underdetermination of the photographic image: what we understand exceeds what we can see. Ruff’s monumental, frontal portrait series seem to embody the paradigm of photographic likeness. Yet they block straightforward readings: the absence of context, pose direction and narrative reduces the visible cues that normally lead to personal attributes. This lack of semantic anchors reveals how interpretation is based on top-down expectations, cultural scripts and situational frames, rather than on visibility alone. Finally, Tillmans’ LED Flicker shifts attention from depicted objects to the conditions of perception. The flickering of LED light and the resulting aliasing effects activate predictive mechanisms that generate micro-hypotheses about rhythm, colour, and continuity. The experience has meaning—rhythmic, affective, epistemic—without showing a stable, recognisable “thing”. Together, these cases show that similarity may be necessary for recognition, but is insufficient for meaning or value. Aesthetic value appears as a context-dependent outcome of generative modelling—where affect, attention, and knowledge acquisition are intrinsically intertwined—and thus reframes classical claims of transparency in a PP-informed aesthetic beyond the merely visible.

Material appearance in stylized representations: the case of design sketches

Billaud Pierre, Bertheaux Cyril, Lavoué Guillaume, Delanoy Johanna

LIRIS

Design sketching is a fundamental tool in product design, enabling designers to rapidly explore and communicate ideas while refining a product’s form and material properties. Design sketches exhibit specific stylistic conventions and non realistic visual shortcuts in order to quickly communicate the idea of a material (e.g. is it glossy or made of metal). Our research investigates how material properties are conveyed in such sketches. We collected 141 hand-drawn sketches of a mug from a pool of both student and professional designers, representing five reference materials. Using these collected images, we then collected perceptual judgements centered on appearance properties such as glossiness and metallicity as well as a realism score. We also asked participants to retrieve the original reference material and match parameters of a material model to each drawing. Our goal is to determine to which extent sketches succeed in conveying the appearance of the target material. Our results reveal that designers had difficulties transmitting the exact same material properties as the reference material, especially for purely diffuse and highly glossy materials; suggesting that it is easier for them to depict “intermediate” materials (somewhat glossy or metallic). Additionally, sketches were more ambiguous regarding the distinction between glossiness and metallicity. Finally, we listed seven visual cues typically used by designers and analyzed their occurrences in the collected drawings. We found that these cues explained a significant part of the judgments. Specular highlights emerged as the most critical predictor, reinforcing the importance of luminance transitions between specular and diffuse light in shaping perceptions.

Material perception in context

Filipp Schmidt, Roland W. Fleming

Justus Liebig University Giessen

Material recognition—distinguishing between substances like metal, fabric, wood, or glass—is essential for predicting object behavior and guiding interactions. Prior studies have shown that people can rapidly and effortlessly recognize materials and infer their properties, despite substantial variability in appearances. However, most previous work used isolated texture patches, ignoring potential effects of visual context. Unlike objects, materials are usually not defined by a diagnostic global shape, rendering them more ambiguous and potentially more prone to contextual influences. Here, we used AI-based image synthesis with careful, iterative prompting to seamlessly embed identical image patches—depicting materials such as eggshell, metal, or water—into eight different surrounding contexts. One context was congruent with the original material (e.g. eggshell patch embedded in the image of an egg), seven were incongruent (e.g. eggshell patch embedded in the image of a human arm). For each stimulus, 10 participants identified the material of the central patch and rated 17 properties (e.g., realism, softness, glossiness, temperature). Importantly, we tested different groups of participants so that no single person saw the same patch twice (irrespective of context). Our findings revealed a strong impact of the context on the interpretation of the central patch. For example, participants identified eggshell as human skin when embedded in the context of a human arm. Property ratings shifted systematically to reflect the categorization—e.g., the patch appeared “softer” and “warmer” in the arm context than in the egg context. A control experiment, in which material patches were paired with mismatched contexts, confirmed that these effects were not due to inattentiveness but genuine perceptual integration of the patch and the surroundings. Overall, the findings demonstrate that material recognition is highly context-sensitive, potentially more so than object perception, highlighting context as a key factor in material perception.

Material Twins, Infrared Strangers: Reconstructing Egyptian Blue in Egypt and the Netherlands

Dina Danish, Jean-Baptiste Maitre

Artist

This talk presents our ongoing artistic research into reconstructing Egyptian blue, the earliest known synthetic pigment and a material known for its infrared luminescence. We set out to remake Egyptian blue from raw materials, both in Egypt and in the Netherlands, as part of a film project. Ultimately, our plan is to directly apply the pigment onto 16mm film to capture its luminescence in motion. To the naked eye, the pigments we produced appear nearly identical. The Dutch version is slightly deeper in tone; the Egyptian version is a bit lighter. Visually, there is nothing that suggests a fundamental difference. But once photographed with a modified camera that records infrared luminescence, the contrast is stark. The pigment made in Egypt produces strong luminescence. The pigment made in the Netherlands produces none. This divergence challenges assumptions about material equivalence and raises central questions about the factors influencing luminescence in Egyptian blue. Are the differences rooted in the chemical composition of the raw materials, firing temperature, kiln structure, or the interaction of ingredients during vitrification? We approach these questions through making, filming, and technical imaging, combining our artistic practice with material science and with experimental archeology. In the talk, we will present footage of the production processes in both countries and comparative infrared recordings of the pigments.

Measuring Directional Appearance of Shiny Fabrics Using an Extended Flop Index

Jiri Filip, Nathalie Roy

Institute of Information Theory and Automation

The flop index is a widely used metric in the coatings industry that quantifies directional lightness variation between specular and off-specular viewing geometries within a fixed in-plane configuration. It describes the perceptual transition between the face color observed near the specular direction and the flop color seen at a glancing angle, characterizing the “color travel” typical of metallic and pearlescent coatings. We extend this concept to the domain of fabrics, introducing a quantitative method for evaluating their directional optical behavior. Directional effects in shiny polyester textiles arise from the combined influence of yarn geometry, reflective properties, and weave structure, which generate local shadowing, masking, and inter-fiber reflections. Unlike coatings, which are largely isotropic under in-plane rotation, woven materials exhibit strong azimuthal anisotropy, causing their appearance to vary significantly with orientation relative to illumination and view. In current practice, apparel designers assess directionality through visual rotation tests under controlled lighting, short drape-and-motion videos, or by using gloss meters and multi-angle color measurements sampled over azimuth. However, these methods yield fragmented indicators rather than a unified quantitative descriptor. We compute the classic flop index across a series of in-plane orientations using bidirectional reflectance distribution function (BRDF) data acquired with a goniometric setup. The resulting directional flop profiles reveal dominant anisotropy directions and relative strengths, capturing subtle variations in sheen and highlight travel caused by weave geometry. The same approach can be implemented using simplified multi-angle instruments outside laboratory conditions. This method establishes a direct link between fabric structure and perceived visual dynamics, offering an objective tool for comparing textiles and guiding digital appearance modeling and apparel design, where controlled directional highlights enhance visual expressiveness, paralleling the coatings industry’s use of the flop index to optimize color travel.

Natural and unnatural representation of gravity in art: an eye-tracking study

Marte Sophie Meessen, Béryl Hilberink-Schulpen

Radboud University

Seventeenth-century Dutch art is known for its emphasis on realism. The realistic qualities include the interrelated gravity-effects of balance, weight, and motion. Examples are the pictured suggestions of running figures and the stable (or unstable) placement of objects on a table. Early modern art theory describes the representation of these effects as crucial to the artwork’s success. But how can the impact of natural and unnatural gravity-effects in paintings be estimated? Thus far, studies in the fields of psychology and neuroscience (Kim & Spelke 1992; Indovina et al. 2013; Lacquaniti et al. 2015; Delle Monache et al. 2021) have found that people (subconsciously) register gravity, and therefore gravitational deviations, in realistic images. In our novel approach, findings and methods from the psychology of perception are applied to examine how gravity-effects are perceived in art. Using eye-tracking, we have studied whether painted details that defy gravitational expectation distract viewers from the most important visual storylines in artworks. The empirical experiment has been conducted in spring 2025 at the Radboud University Nijmegen. The study combines art-historical analysis, with quantitative methodologies and a qualitative inquiry. It examines how naïve participants distribute their attention across a selection of Dutch seventeenth century paintings from the Rijksmuseum Amsterdam collection, specifically when there is a (photoshopped) gravity-defying-detail in the composition. The findings of this interdisciplinary study will be presented and discussed at the conference, as well as the added value of empirical methodologies to answering the art historical questions on perception and embodied cognition.

Novi Mundi, A Journey of Discovery Through Uncharted Human Realms

Persijn Broersen, Margit Lukacs

Broersen & Lukacs

In this talk, we explore how objectivity emerged in the observation of natural phenomena and how scientific ideals and everyday working practices became intertwined in the creation of scientific imagery. We aim to illuminate shifting perspectives, ideological undercurrents, deceptive visual conventions, and the seemingly objective character of scientific depictions of nature. The work we use to clarify these ideas originates from a commission from the Radiology Department of the Leiden University Medical Center (LUMC). For this project, we took the famous 18th-century anatomical drawings of Albinus from the Encyclopaedia Tabulae Sceleti as a conceptual point of departure. This led to a series of landscape-like images developed through advanced radiological observation techniques. These works use models of human organs both as objects of clinical scrutiny and as terrains onto which viewers can project narrative, desire, and imagination. In these large-scale “landscapes,” empirical data and objective measurements blend with subjective and fictional elements. Using 3D models derived from medical imaging, we repeatedly tilt the camera— and thus the gaze— to generate new horizons that unfold like terrains awaiting discovery. Since prehistoric times, the human brain has been attuned to scanning horizons for irregularities; such disruptions sharpen perception, while regularity lulls it. By constructing unfamiliar vistas from scientific observations, we fuse fact with fiction to activate imagination and heighten observational awareness. These hybrid images invite dialogue and collective reflection, encouraging viewers to consider the links formed in the act of observation—between the empirical and the remembered, and between what is observable now and what is possible in the future. By encountering these images from renewed perspectives, we open pathways for new connections and new ways of seeing, where multiple worlds converge.

One Motif, Four Media: What Happens to Art Experience Across Medialities?

Hanna Brinkmann, Alexandra Victoria Alvarez, Luitgard Voller, Eva Specker, Matthew Pelowski

University for Continuing Education

The inter- and transdisciplinary research project „Art Experience in the (Post-)Digital Age {original | digital | virtual}“ (OrDiV) investigates how art engagement differs across media formats—analog originals, digital reproductions, augmented and virtual environments. Against the backdrop of accelerated digitization and the proliferation of immersive technologies, the project examines how shifting media ecologies influence attention, cognitive and emotional involvement, and the after visit memory of the art works. Building on theoretical debates surrounding aura (Benjamin 1936), the status of the digital image, and recent museum practice, OrDiV embeds its empirical design within broader discussions of art experience, depiction, and mediation. Concepts from the Vision and Depiction Framework (Wijntjes & van Middelkoop 2024)—including formal analysis of texture, colour/light, space, and material; the distinction between medium and motif; pictorial space; and semiotic/iconographic layers—inform the interpretation of comparative reception data. Across four studies, museum visitors at the Upper Belvedere in Vienna encountered seven works of Austrian Expressionism under different media conditions. Using mobile eye tracking, qualitative interviewing, online questionnaires on cognitive–emotional response, and a one-week post-visit online survey, the project captures multimodal indicators of visual attention (fixations, gaze mapping), meaning-making, and memory. The analysis integrates top-down defined areas of interest, automatic event classification, thematic coding (Braun & Clarke, 2006), and measures of art interest and knowledge. Complete datasets include 55 participants for the original paintings, 63 for digital reproductions, 50 for the VR-twin, and 64 for the augmented reality application. The proposed conference contribution presents the project’s methodological framework and first comparative findings, highlighting media-specific qualitative differences in art experience. By combining art-historical, psychological, and media-theoretical perspectives in a mixed-methods design, OrDiV contributes to Digital Humanities, empirical aesthetics, and ongoing debates on how mediation shapes art experience in the (post-)digital age.

Pictorial forms in movement

Valentine Bernasconi

University of Bologna

The relationship between words and images, words and forms, has never been as strong as it is in the era of AI. Large Language Models connect the two, shaping a latent space that is difficult to imagine from a human perspective. These models can be used to automate the description of images, paintings, or forms. Moreover, by concatenating different generative models, it becomes possible to animate depicted characters and make them speak—the so-called “speaking Mona Lisa,” in which the face of one of the most famous women of the sixteenth century begins to move and express herself. From the digitised form emerge words, transformed into sound, and then into moving forms—a radical shift in paradigm. These profound operations on paintings—made possible because they are digitised and placed within the latent space—challenge traditional interpretive frameworks in art. From a phenomenological perspective, they interrogate how digitisation and new technologies revolutionise the way paintings can be perceived, as well as the new kinds of information they offer to contemporary viewers. Yet, earlier artistic experiences in religious contexts show that forms were sometimes also meant to move. Originating from the Word, religious images could animate to shape a spiritual experience. Thus, artworks were not necessarily created solely for the intellectual mind but for the people, and interpretations did not always take the form of verbal descriptions. Through an experiment that makes artworks “speak” based on generated descriptions and generative tools, this paper aims to examine more deeply the impact of digital transformations on the construction of knowledge about art, the contemporary bond between words and forms, and on the new modes of representing artistic forms within digital space.

Presenting IRECONA: An Interdisciplinary Method for Appearance-Based Reconstruction of Degraded Color Field Paintings

Stefanie De Winter, Willemijn Elckhuizen, Pascal Barla, Francois Margall, Morgane Gerardin, Koen Brosens, Geert Van der Snickt and Johan Wagemans

KU Leuven

Many postwar paintings have undergone significant degradation. Due to their atypical structure, often lacking a gesso or varnish layer, and the use of industrial paints such as fluorescent DayGlo pigments, they are highly susceptible to ageing. The unprimed canvas, frequently left visible within the pictorial surface, is also prone to oxidation. This degradation is irreversible and affects the very core of these works’ meaning. Unlike traditional painting, where pictorial meaning can exist independently of the material substrate, the meaning of these works resides precisely in the visual effects that emerge from their specific material composition. Over time, these works have been silently transformed, obscuring their original perceptual impact and leading to misinterpretations in art-historical discourse. The Interdisciplinary Reconstruction of Art (IRECONA) project addresses this issue through appearance-based-reconstruction. Drawing on archival documentation, chemical analyses of original materials, and artificial ageing studies, the material composition and degradation of two case studies, Frank Stella’s Effingham I and Morris Louis’s Alpha Sigma, were examined in depth. These empirical data form the foundation for an AR-interface that allows viewers to navigate between the artworks’ presumed original state and hypothetical future stages of decay. An upcoming perception study at the Van Abbemuseum (February–March 2026) will combine eye-tracking, questionnaires, and interviews to examine how (non-)experts experience these reconstructions in front of the real artworks. Participants explore the reconstructions using a Varjo XR-4 headset. One group receives contextual information about material ageing, a second views a control video on abstract art, and a third views the original paintings with a mobile Tobii eyetracker, also after seeing contextual information. We hypothesize that informed participants will fixate longer on unstable zones, report greater awareness of perceptual change, and show increased aesthetic appreciation.

Process-Aware Robotic Painting: A Closed-Loop Framework for Human–AI Collaborative Creation

Dantong Qin, Alessandro Bozzon, Qinlin Liu, Yike Guo, Pan Wang

Delft University of Technology

Enabling AI to respond appropriately to human input in turn-based collaborative painting remains a major challenge. Most existing systems operate in a single-turn or reactive manner, lacking process-level reasoning, which results in inconsistent visual progression and reduced interaction fluency over time. Achieving a responsive collaboration requires the AI not only to generate visual content, but also to perceive the evolving canvas, interpret human intent, and plan its next action within a continuous creative process. To address this issue, we propose a process-aware painting system that allows a human and a robotic arm to paint alternately on the same canvas around a shared theme without additional manual operation. The system forms a closed perception–reasoning–execution loop that enables iterative coordination in a physical environment. It consists of three integrated modules: a vision–language planner that performs semantic and spatial reasoning to predict what to paint next and where; a painting synthesizer that refines the target region using multimodal and visual models to produce grounded, editable responses; and a stroke-based executor that converts these predictions into human-like brush trajectories for robotic painting. Through this design, the AI can respond dynamically to human strokes, maintain contextual coherence, and contribute to the artwork’s progressive development. Experiments demonstrate that the proposed system more accurately predicts user intent and generates temporally consistent updates across multiple turns, outperforming reactive baselines. The framework provides a computational foundation for process-level collaboration between humans and embodied agents, illustrating how language and generative models can be embedded into real-world painting workflows to achieve adaptive, physically grounded visual generation.

Reconstructing the Imagetic Brain: Speculative Design and Visual Image Reconstruction of Closed-Eye Visualisations

Ana Domingues

Goldsmiths University

This paper investigates closed-eye visualisations (CEVs) as a fertile ground for interdisciplinary formal analysis, situated at the intersection of art and scientific innovation. Building on my doctoral and MA research, I focus on the formal elements of subjective internal imagery—texture, color, motion, and light—arising during phenomena such as migraine aura, 5-MeO-DMT, and ayahuasca experiences, and explore their expressive potential through digital animation.

Central to this inquiry is the transformative context of externalisation technologies. Recent breakthroughs in brain imaging, particularly the work of Nishimoto et al. (2011), and the Kamitani Lab (2017), now enable the reconstruction and externalisation of visual experiences from neural data, making it possible to glimpse images produced internally by the brain without external visual input. Simultaneously, the rapid emergence of computer vision technologies is reshaping the analysis, synthesis, and dissemination of visual information across creative and scientific domains. These twin advances foreground not only new possibilities for depiction, but also raise important questions about authorship, agency, and ethical responsibility in the capture and interpretation of inner imagery.

My artistic practice features a series of generative animations translating subjective visual experiences into dynamic audio-visual forms. Merging phenomenological interviews and computational techniques, the research interrogates the interplay between motif (representation) and medium (surface), and examines how formal elements act as both perceptual structures and carriers of meaning.

By situating artistic practice within the evolving landscapes of externalisation machines and computer vision, this work offers a speculative framework for the future of visual expression, therapy, and analysis. It contributes to critical conversations about the powers and responsibilities involved in visualising the invisible, and the creative, therapeutic, and ethical frontiers opened up by these technological advances.

Spilling the Tea Ceremony: On Becoming DATA DIVA

Antye Guenther (baby) DATA DIVA

Maastricht University / Jan van Eyck Academie

Antye Guenther, aka (baby) DATA DIVA, is a glitter-loving knowledge-inventor and unreliable narrator, born in a country that no longer exists. Rumoured to have been a former child test subject in Soviet brain experiments, she now joyfully interrogates the conditions and fictionalities of Western knowledge and data regimes. With a warm fondness for troublemaking, she mischievously crosses disciplinary boundaries, often while crafting cheap-ass rhinestone jewellery as part of an ongoing collaborative glitter-as-(communal)-care-practice.

This keynote invites the audience into the evolving processes behind the emergence of DATA DIVA, a fabulous alter ego for doing artistic research otherwise. DATA DIVA collaboratively examines biometric data visualisations, with a particular focus on brain imaging software, tracing the hierarchies, norms, and ideologies embedded in contemporary 3D imaging practices. By closely observing how digital (rat) brains materialise, DATA DIVA asks how these technologies actively shape our (non)understandings of brains and bodies. Engaging with powerful tech-science knowledge regimes that present themselves as natural, precise, and self-evident—while elegantly obfuscating the decisions, assumptions, and value systems built into interfaces, automated processes, and algorithms—she focusses on using non-digital tools to critically examine these digital regimes. Truly invested in what might be called Deep Superficiality Studies, the keynote follows DATA DIVA as she indulges in slick, seductive surface meshes representing volumetric MRI brain data; catwalks the neuro-lab in her custom DD lab coat; and develops an unapologetic, performative voice in conversation with researchers, neuroscientists, and data enthusiasts alike.

Study 7/0: Locus Erroneum

Dejan Grba

Interdisciplinary Graduate Center, University of the Arts, Belgrade

In Study 7/0, we leverage GPS errors as a source of generative emergence to indicate the tensions between navigation technologies and their users’ individual directions and interests. We gathered the error data by running a TrackLog function on a fixed GPS receiver for seven days nonstop. Instead of a single location it should have ideally registered, our TrackLog captured 8438 distinct positions, indicating motion on a 35km trajectory that covered a 2 square kilometer area with an average speed of 200m/h. This is a combined consequence of the GPS device’s operation under changing weather conditions and the GPS infrastructure’s intrinsic limitations. Using the longitude and latitude of the collected waypoints and proportionally adjusting their original timestamp deltas to reach five minutes of total running time, we animated a red dot along the horizontal projection of this wandering path. We contextualized the visualization with a synchronous display of all TrackLogged values (altitude, heading, distance, date, speed, etc.). To generate sound, we treated the virtual, data-implied movement as a traversal of a listening device in real space with the ensuing acoustic phenomena: Doppler effects and volume modifications. The resulting installation comprises a large-scale video projection and a two-channel audio. While common applications of navigation instruments such as GPS rely on accuracy, we focus on imprecision as an inverse but equally revealing index of their usefulness. The interplay of shape, space, and signs in our three-dimensional rendering of a four-dimensional entity indicates the necessary yet problematic information collapse in mapping and wayfinding representations. GPS tools calculate routes as tradeoffs between environmental constraints and desired accessibility, so their slipups incur practical damage and, in a broader sense, also highlight the mismatch between our diverse personal intents and our technologies’ pragmatic but often coercive functional requirements.

Style Needs Time: Temporal Dynamics in the Recognition of Stylistic Differences in Painting

Anna Miscenà, Jozsef Arato, Gregor Hayn-Leichsenring, Corinna Kühnapfel, Raphael Rosenberg

University of Vienna

The ability of identifying an artistic style is generally understood as a trained skill, allowing art historians and connoisseurs to classify a painting or estimate its period from visual analysis alone. Yet the perceptual mechanisms underlying this recognition remain insufficiently understood. The present study investigates how viewers engage with stylistic information at eye-movement level, and how this engagement unfolds over time. We conducted an eye-tracking experiment in which participants viewed paintings categorized along two dimensions: (1) stylistic group (realistic vs. impressionistic) and (2) quantified low-level visual features, including HOG-based complexity, self-similarity, anisotropy, and colorfulness. Mixed-effects modelling revealed that fixation duration for the full viewing time (8 s) was significantly modulated only by visual complexity, with no robust effect of style. However, when fixation patterns were analyzed dynamically over time, style emerged as a significant predictor. A divergence between stylistic categories appeared approximately 2.5 s into viewing, indicating a temporally specific shift in attention that cannot be explained by low-level features alone. We propose that this delayed divergence reflects cognitive effort associated with learning or accessing a “vocabulary of style” - a set of visual associations that are more complex than simple low-level image properties. Qualitative data from an open-response questionnaire support this interpretation. Finally, this study situates its empirical findings within the context of formalist art history, suggesting that examining stylistic processing at a physiological level offers several advantages for the concerns of contemporary vision science and art history. Understanding how viewers learn to recognize style over time may inform the training of connoisseurship and perceptual expertise, support media-literacy education, and provide broader insights into how complex visual categories are formed and refined in the future.

Technics, Individuation, and the Limits of Categorisation: A Stieglerian Intervention in Palaeolithic Archaeology

Karel Kuipers, Vincent Niochet, Tullio Abruzzese

Leiden University

Archaeology started as a science based on categorisations. Archaeological material cultures are defined based on the re-occurrence of artefacts’ visible physical properties. However, this reasoning often results in static, reductive and essentialist categories that explain little about the socio-cultural processes that led to their creation and shaping. In other words, despite numerous theoretical advancements in the field, archaeology in some cases has still not managed to shake off its cultural-historical feathers. In the study of the Palaeolithic period (ca. 3.3mya-11.5kya), this “categorisation logic” led to the creation of static categories (archaeological cultures). While alternative approaches to these categories exist, they have so far not succeeded in displacing this paradigm in archaeological research. The recognition and understanding of cultural processes, such as innovation, transmission and loss of cultural traits, are hindered by the reductive and essentialist nature of those archaeological cultures. As an alternative, we propose approaching the issue through Bernard Stiegler’s process-based concept of transindividuation. This concept entails the idea that people and groups form their identities through the exchange of technical and cultural expressions. Individuals, technical objects, and communities recursively shape and constitute each other. To illustrate possible outcomes of the proposed approach, we will attempt to apply Stiegler’s framework on an Upper Palaeolithic case study, demonstrating that if we analyse artefactual visual characteristics as emergent traits rather than a series of attributes that should be fitted within a narrative, it is possible to try and infer inter-personal contacts and cultural exchanges previously hidden. While this approach cannot solve all conceptual tensions in Palaeolithic archaeology, Stiegler’s ontology of technics does offer a precise diagnostic tool to clarify where archaeological analysis falls short regarding the dynamic processes that shape technical forms and collective and individual life.

The agency of form and material: from bubble wrap to art

Arthur Crucq

Leiden University Centre for the Arts in Society

For centuries, art philosophy has reflected on form and its essence. For Immanuel Kant, the aesthetic judgment involved the disinterested judgment (reflection) of the pure form of natural and artistic objects independent of material qualities. However, if we approach form as the result of creative processes and consider the verb ‘to design’ as denotating such processes, the significance of “sign” stands out. Design as a human activity can thus be understood as generating signs and as such as instigating processes of meaning-making. In this talk I would therefore like to depart from form as de-sign and consider de-signs as accumulations of (artistic) motifs which are in themselves already signs. I will make clear that prior to be attributed with content, it is through their formal and material properties that motifs in art already incite processes of meaning making. I therefore use an understanding of the concept of meaning beyond meaning as the representation of concrete content. After all, signs trigger bodily responses. Such responses can be explained from what the de-signs’ form and materials afford. As example, I will begin by discussing the pleasure gained by such ritual and auditory sensations like the puffs generated by the repetitive squeezing of bubbles in bubble wrap that is, as it were, offered as a possibility by the material properties and de-sign of the material. I will argue that artists and designers consciously and unconsciously make use of such affordances. Therefore, to understand how processes of meaning making unfold I argue that form must be considered as de-sign(s) operating both within semiotic networks (sign systems) and social networks of agents within which (artistic) form, their constitutive (sensorial and material-) motifs, thus de-signs, function themselves already as agents acting upon, and on behalf of, others.

The colour of iridescent samples at every moment of a bumblebee’s flight

Li Shiwen, Hannah E. Smithson

Oxford University

Iridescence is a dynamic visual phenomenon whereby colour changes with viewing angle. It is commonly found in nature (e.g., beetle carapaces, bird feathers, and flowers pollinated by bees). Because the observer can change the appearance of an iridescent object by simply looking at it from a different angle, iridescent stimuli are particularly suitable for studying active perception. Using bumblebees (Bombus terrestris) as our model organism, we introduce a novel experimental paradigm that captures exploration and behavioural interactions with visual stimuli.

Although iridescence has been suggested to serve various signalling functions in nature, we still lack an understanding of the specific percepts that separate “iridescent” from non-iridescent surfaces. Conventional psychophysical methods collect categorical responses, which are limited such that they collapse the spatial (colour variation across the surface) and temporal (variation over time with movement) dimensions of perceptual variation to low dimensional responses. The visual appearance changes are contingent on viewing geometry, therefore, characterising dynamic behavioural exploration of visual signals may reveal signatures of material depiction that are relevant to the perceptual agent.

We present firstly a stereo-tracking pipeline to extract bee flight coordinates and reproduce the visual signals available to the bee at every moment in flight. We then describe a method for quantifying how bees dynamically sample iridescent stimuli. Finally, we present an experiment in which bumblebees learned to discriminate between horizontal and vertical bars. Even though the material properties of the stimuli (colour and presence of iridescence) were irrelevant to the task, we found that bees preferentially hovered at locations where minute shifts in position would lead to perceivable changes in surface colour. We discuss how our methods and results can be used to reveal relevant percepts for efficient communication of material properties, both in biological contexts and technological applications.

The Influence of Iconographic and Formalistic Instruction on Art Viewing Behavior

Yuexi (Betty) Zhu, Zsofia Pilz, Maarten Wijntjes

Utrecht University

This eye tracking study investigates how iconographic versus formal art viewing instructions influence gaze behavior during the interpretation of artworks. Iconographic analysis emphasizes symbolic content, narrative meaning, and cultural symbolism, while formal analysis focuses on perceptual properties such as color, light, texture, and space. Participants were randomly assigned to one of two instructional groups and watched a short video introducing the corresponding analysis method. They then viewed ten paintings, during which their eye movements and verbal descriptions were recorded. Areas of Interest (AOIs) were defined and categorized as iconographic or formalistic for each painting. The result revealed a significant interaction between AOI type and instructional group, indicating that participants fixated longer on AOIs consistent with their assigned group. Additionally, participants in the formal analysis group made more saccades overall, suggesting a more exploratory viewing strategy. Verbal analysis revealed that participants in the iconography group used interpretive and symbolic language, while participants in the formal group described visual elements in a more technical and observational manner. These findings demonstrate that top-down instructional context significantly shapes visual and cognitive engagement with art. Museums and art galleries could enhance visitor experience by tailoring interpretive materials, such as labels, guided tours, or digital materials, to promote specific types of art engagement based on educational goals or visitor interest.

The Shape of Gender: How Context Becomes Form in Artistic Representation

Eftychia Stamkou, Selina Khan, Matteo Tafuro, Nanne van Noord

University of Amsterdam

Museum collections have historically, and continue to, consist predominantly of artworks produced by male artists (Heithuis et al., 2022; Christensen, 2016). This male-dominated artistic canon has profoundly influenced not just who creates art, but also what is depicted and how it is visually constructed. Systematic marginalisation of female artists may itself influence the development of distinct artistic tendencies by shaping access to training, materials, subjects, and audiences, producing gendered patterns both in form and content without assuming essential differences (Nochlin and Grant, 2021).

Complementing this structural account, gender role theory holds that societies develop distinct norms regarding how different genders should behave, appear, and function Eagly and Wood (2012). These gendered roles in turn surface in visual culture as stereotypes that shape depiction, for example, in emotional expression (Dickey, 2010), professions and social settings (Thiel, 2016), and sexualisation (O’Kelly, 1980). This raises two related questions: Do male and female artists differ in how they portray their subjects? And is there such a thing as a gendered artistic style?

To investigate these questions, we compiled a large-scale dataset of over one million artworks by 45,000+ artists, dating from 1800 BC to the present, enriched with metadata on artists (e.g., gender, nationality, birth-year, years active) and artworks (e.g., medium, style, textual description of content). Using Multimodal AI and statistical analysis, we examine how the artist’s gender relates to both low-level visual features (form) and high-level semantic content (e.g., stereotypical portrayals, subject matter), uncovering systematic biases in how women are represented in visual art.

The Softness of Screens: Liquid Visual Culture in Contemporary Displays

Beatrice Sartori

University of Bologna

This visual research investigates the soft materiality of screen-based technologies by tracing a genealogy that links their contemporary flexibility and responsiveness to the hybrid behavior of liquid crystals, the foundational material of modern displays. The study foregrounds the dual nature of liquid crystals, organic materials that flow like liquids while exhibiting crystalline optical anisotropy, as key to understanding how screens became mutable and intelligent media. Their phase transitions, responsiveness to electric fields, and ubiquity in biological systems demonstrate an enduring material vitality that continues to shape the sensory and operational capacities of screen-based interfaces. Alongside this material investigation, the paper examines the emergence of a visual culture of liquidity that developed in parallel with the evolution of display technologies. Case studies include the early 2000s Y2K and Frutiger Aero interface aesthetics, characterized by gloss, translucency, and biomorphic refractions that expressed a cultural fascination with fluid surfaces; early screensavers that staged bubbles, waves, and iridescent flows as demonstrations of LCD optical precision; and the Nokia Morph (2008) concept, which envisioned stretchable, self-cleaning, and adaptive devices inspired by biological materials and nanostructures. These imaginaries resonate with current research on interactive deformable displays that use liquid metals as conductors (Miyakawa et al., 2025), showing how representations of liquidity anticipate and shape technological development. This continuum also informs the visual languages of post-internet media art, where glossy, fluid, and soft interfaces function both as aesthetic resources and conceptual frameworks. Building on Laura Tripaldi’s idea of soft technologies (2025), the paper proposes the screen as a prosthetic sensorium whose material behavior is inseparable from its epistemic and aesthetic functions. Understanding the liquid interface, at once material and cultural, is therefore essential for interpreting contemporary media art and for recognizing today’s displays as hybrid and agentive media.

This is not an AI generated title.

Flip Phillips

Rochester Institute of Technology

Drawing and painting gave us the attempted imitation of the visual world. Photography promised its capture. Computer graphics exposed this promise as an illusion. And now AI completes the circle, revealing that every image, still or moving, has always been a fabrication, just made with different instruments. From the caves of Lascaux, images have been about the human desire to communicate, to tell stories, to share visions of reality or, more importantly, possible ones. The tools have changed, but the impulse remains the same. Hand in hand, the tools have changed us as well. What is there to believe about depictions – Seeing is no longer believing.

Currently, there is a widespread anxiety both about the authenticity of images and about the potential for AI to replace or destroy the artist and/or the creative process entirely. Throughout the history of art, new technologies have been met with similar fears. Photography was once thought to be a threat to painting, yet it ultimately expanded the possibilities of visual expression. As a nominal pioneer in computer graphics, I have witnessed firsthand how its technology was, at first, perceived as an existential threat to traditional art forms, only to become a new tool and medium for creativity and innovation.

Like computer graphics, generative AI is a force that will ultimately be integrated into the process of visual creation. Furthermore, it is silly to assume that AI has not already been used in the creation of images, static and moving, for decades. The ongoing challenge is to understand and harness the evolving power of these tools to collaborate with human creativity rather than to compete with it.

This talk will examine current uses of technology that are in the spirit of collaboration rather than competition, especially with respect to motion pictures and their production.

Through a Child’s Eyes: Exploring Gaze Patterns on Albers’ Homage to the Square

Bade Kilic , Zsofia Pilz , Dr. Francesco Walker, Prof. Dr. Mariska Kret

Leiden University

Children learn about art through exploration and personal interactions with artworks. This makes museums potentially rich environments for learning and development. Despite the creative and imaginative experiences that abstract art offers, few studies have focused on how children engage with abstract art and how museums can encourage this interaction. Understanding children’s responses to abstract art can guide museum practices specifically targeted for children, encouraging their aesthetic curiosity and creative knowledge development (Piscitelli & Anderson, 2001). This study examines how interactive elements influence children’s engagement, preferences, and emotional responses to abstract artworks, analyzing three paintings under three different conditions: red-tinted glasses, color paper interaction, and free viewing at the Albers Museum in Germany. Each interactive condition alters the perception medium through changing the relationship between color form and perception, thereby creating a unique experience for each painting. Three paintings from Joseph Albers’ Homage to the Square series were selected, which share identical geometric properties and format and only differ in colors. These specific qualities of the paintings allow for neutral comparison of the interactive elements’ influence on engagement and emotional responses, independent of compositional variation. The main goal is to determine whether interactive elements lead to greater engagement and preference for the paintings. Engagement was quantified using eye tracking, with a focus on gaze patterns within predefined areas of interest for each painting. Children’s emotional responses to the artworks were captured using a questionnaire that assessed aesthetic appreciation and preferences. The study hypothesizes that participants viewing a painting with interactive elements will report higher engagement compared to those with no interaction. This study emphasizes the potential benefits of incorporating interactive approaches into museums to help children engage more deeply with abstract art.

Visions of Veins: Eye-Tracking the Irregular Symmetries of Book-Matched Marble

Xingyu Long, Teresa Kamencek, Raphael Rosenberg

University of Vienna

Book-matched marble, i.e, mirrored slabs cut from a single block, offers a unique opportunity to study how material surface organizes visual exploration without recourse to depiction. Building on the “amimetic aesthetic” of marbling (Rosenberg 2021), which links its appeal to both the suggestive discovery of figures and the pleasure of patterned variety, we approach symmetric marble—or, using an art-historical neologism, “opus sectile symmetricum” (Rosenberg 2025)—as an interface between medium and motif, form and perception. We present eye-tracking results comparing photographs of mirrored and non-mirrored polychrome marble slabs, with viewers inspecting each image for 30 seconds (Brodnik 2025), and analyze spatiotemporal gaze patterns and post-trial aesthetic judgments. Four tendencies emerge: (i) structure guides attention: on plain slabs, saccades trace mineral veins, while on book-matched slabs, the gaze first localizes the axis and then tests symmetry relations; (ii) viewers explore vertical symmetries more than horizontal ones, consistent with habituation to bilateral bodies and faces; (iii) time courses invert canonical viewing: rather than shifting from long to short saccades, symmetric slabs elicit increasing amplitudes as observers range across ever-wider correspondences; (iv) natural, irregular symmetries sustain engagement and are preferred to flawlessly artificial ones. These findings quantify how irregular symmetries transform a variegated surface into a cognitively rewarding search space. Art-historically, they help explain the enduring appeal of book-matching from ca. 100 AD to the present, showing why architects and patrons invested in large, veined stones whose mirrored placement amplifies order within disorder. We thus propose book-matched marble as a natural laboratory for Vision and Depiction: a case in which the physical medium itself orchestrates visual inference and aesthetic experience.

Who Appreciates Ambiguity in Art?

Rajat Ravi Rao, Francesco Walker, Manon Mulckhuyse, Michiel van Elk

Leiden University

Background: Semantic instability (SeIns) is the recursive stabilisation and destabilisation of interpretations while viewing art which shapes one’s aesthetic experience. Existing theory predicts higher interpretive divergence enhances appreciation with individual differences potentially being relevant in determining the relationship between appreciation and ambiguity. However, how semantic properties interact with visual features remains unclear. Aims: We investigated how semantic distance and insight-confidence predict aesthetic appreciation and predicted them to have a negative correlation. We predicted semantic distance would positively correlate with interest, emotional intensity, and cognitive affect, whilst insight-confidence would negatively correlate with the aesthetic appreciation measures. Ambiguity tolerance was predicted to moderate these relationships. We additionally examined whether visual image properties moderate semantic distance and insight confidence effects. Methods: n=123; rated interest, emotional intensity, and cognitive affect for 60 artworks selected via computational visual diversity sampling. Semantic distance was operationalised from interpretation embeddings (s-BERT). Insight-confidence measured participants’ confidence in their interpretations. Image features were extracted using the Aesthetics Toolbox. Multilevel regression models examined how semantic distance, insight-confidence, and image features predict aesthetic appreciation, and how ambiguity tolerance (MSTAT-II) moderates these relationships. Concurrent eye-tracking data are being collected from a different set of participants to investigate the relationship between semantic distance and visual exploration patterns with ambiguity tolerance as a moderator. Results: Contrary to predictions, semantic distance negatively predicted all aesthetic measures. Insight-confidence independently predicted outcomes more strongly. Visual properties independently predicted aesthetic responses but did not moderate semantic distance effects. Ambiguity tolerance significantly moderated interest only. Conclusions: Semantic and visual properties operate independently on aesthetic appreciation. Interest is personality-mediated whilst emotional intensity and cognitive affect operate through trait-invariant mechanisms. Eye-tracking analysis will clarify whether visual exploration patterns are correlated with computational measures of ambiguity.

Seurat’s dots and the myth of optical mixing

Jeroen Stumpel

Utrecht Univerity

In 1886 Georges Seurat presented his large painting A Sunday Afternoon on the Island of La Grande Jatte at 8th (and last) Impressionist Exhibition.
Instead of an impressionistic and fluid painterly touch, the canvas was filled countless stripes and dots. The technique came to be called pointillism, although Seurat himself preferred the term divisionism. It is well known that Seurat had studied modern theories of perception and colour, in particular those of the French chemist Michel Eugène Chevreul.

The use of dots is always related to the idea of optical mixing, implying that Seurat used his miniature dots to blend his colours not on the palette, but on the retina of the viewers. His technique supposedly aimed at additive, rather than subtractive mixture. But I believe this view ignores the actual pictorial problem Seurat dealt with, as well as the particular background of Chevreul’s work.

Chevreul worked as director for the Gobelins Manufactory in Paris, which had as its main task the production of tapestries, a form of decorative art, and published his
views in 1833. One of his main ideas was that there is a fundamental difference between techniques that made use of infinitely fine gradations (such as oil painting), and those that had to work with larger, separate patches of colour, such as fresco, mosaics, embroidery and of course tapestry. Fifty years later, by the time Seurat began to experiment, traditional oil painting with its fine, invisible blending of colours had become suspect. This was part of a new, decorative aesthetic, that held a negative view of the autonomous easel painting, which often relied on refined illusionistic techniques. Seurat starting to use his dots not to obtain optical mixture, but to paint in a technique that he considered decorative, in opposition to the illusionistic technique of traditional oil painting. Inspired by Chevreul’s ideas on tapestries, he began to paint like a tapestry weaver. Rather than wishing that his viewers would, from a certain distance, optically blend the dots away into more bright and purer colours, he wanted his dots to remain distinct and visible. He even adapted his pictorial subjects to themes that had been advised by Chevreul for decorative work.

Natura Pictorum: Bruegel’s Macchia and the Ecology of Landscape

Joris van Gastel

Universität Leipzig

In his seminal yet controversial 1934 essay on Pieter Bruegel the Elder, Austrian art historian Hans Sedlmayr draws a strict distinction between the painter’s depictions of people and nature. Even if these worlds converge in Bruegel’s later works, according to Sedlmayr “the tear itself remains tangible in the picture: in the formal sphere as a hidden dissonance, in the objective sphere as the demotion of the human to the animal or vegetable.” This contribution will examine what remains of this distinction as well as Sedlmayr’s broader formal analysis of Bruegel’s oeuvre when approached not, as Sedlmayr does, from the perspective of the depicted figures, but from the representations of nature. In doing so, the argument builds upon recent ecological approaches in art history, with the aim of closing Sedlmayr’s rift and repositioning the Bruegelian human in the painter’s artistic ecosystem.