Art and vision

Drawing expertise predicts not just quality but also accuracy

Linda Carson, Matthew Millard, Nadine Quehl and James Danckert

Drawing is complex. When people draw representationally from close observation of a subject, the completed drawing is a product of many factors: their visual perception, analysis and knowledge of the subject; their ability to translate that information to the page; their judgment of which properties of the subject and of the drawing are most important to success; and, their ability to detect and correct errors in the drawing as they go. In order to tease apart the successes and failures of these contributing factors, we developed and validated objective measures of local and global drawing accuracy. We analyzed shapes in the image as polygons with vertices at unambiguous landmarks and measured four independent dimensions of drawing error of those polygons: orientation, position, scaling and proportionality. We collected 34 drawings of a complex still life and over 100 qualitative ratings of those drawings. The ratings and overall polygon accuracy were consistent with the drawing expertise of participants, and experts were more accurate than novices on every dimension of polygon error. We discuss prospects for the validation and improvement of drawing training, and how these measures can contribute to the study of visual perception.

The specificity of art revealed by an empirical study on multi-cultural aesthetic material

Jana Boehringer, Claus-Christian Carbon and Stella J. Faerber

Research in the field of empirical aesthetics often uses definitions, theories and experimental paradigms for aesthetics quite arbitrarily, especially by using artworks as just one typical class of aesthetically relevant material among other comparable domains. We demonstrate the exclusiveness of visual art by analyzing the specific relationships between key variables of aesthetics in different domains. Participants evaluated material from five visual categories (artworks, costumes, faces, houses, landscapes) originated from four different countries (Austria, Germany, Japan, Tanzania) on two main variables associated with aesthetic appreciation - familiarity and liking - and classified the material's country of origin. Familiarity and liking were only related for artworks (R>.85) demonstrating a general preference for art being familiar, an empirical pattern not observable for the other classes of material.This indicates that preferred artworks should not be accompanied with too much uncertainty and arousal. Even more interesting, when analyzing the data on a stimulus level, specific artworks did not benefit when being evaluated as familiar vs. unfamiliar, but strongly gained liking when the material was correctly classified. This specific pattern was observed for artworks but not for any other class of material, underlining the importance of successful classification, for the appreciation of the arts.

What is the main ingredient for transforming an ordinary object to a piece of art? Aesthetic Evaluations on objects of judgments ambiguous art quality.

Manuela Härtel and Claus Christian Carbon

There is a fast growing body of empirical studies on aesthetic appreciation and aesthetic judgment, particularly on visual art. It is, however, still under debate what main ingredient is needed to identify an object as a piece of (modern) art. We know from recent questionnaires studies on aesthetic judgments that visual art is tightly linked with variables such as stimulus-symmetry, complexity, familiarity, artistic style, appeal to social status, and personal preferences [e.g., Augustin et al, 2012, Acta Psychologica, 139(1), 187-201]. We specifically aimed to find out which variables are mainly accountable for creating artistic quality of objects of ambiguous aesthetic quality. To cover a wide spectrum as possible, we utilized 213 objects in total and asked persons who were untrained in art perception to assess the objects variables commonly summed to be linked with artistic quality. Out of the used variables (preference for the objects, originality, ambiguity, understanding of the objects), the variable originality served as the best predictor for artistic quality, whereas preference was found only of minor importance. Additionally, high artistic quality was often related to higher degrees of ambiguity and lower degrees of understanding.

How should I place two rectangles so they look best?

Françoise Samuel and Dirk Kerzel

We observed that in the absence of symmetry, equilibrium does not explain, or only to a small degree, whether a composition is considered aesthetically pleasing. We define equilibrium in visual compositions in an analogous way as in physics: dark areas represent weights and a composition is equilibrated when the center of mass of the dark areas is in its center. Although the position of the center of mass of the composition has nearly no impact on the ratings, does it interact with other properties of the composition in order to influence the aesthetics ratings? We varied the shapes and the area ratios of the two rectangles making up the compositions, and also the position of the center of mass of the compositions. Our results show an interaction between the position of the larger rectangle and the position of the center of mass. That is, for a given position of the center of mass either on the left or on the right, compositions in which the larger rectangle is on the side of the center of mass are preferred over compositions in which the smaller rectangle is on the side of the center of mass.

Perspective Errors Masked by Architectures in Paintings

I-Ping Chen and Chih-Hsiang Lin

Ever since the discovery of the principles of linear perspective, Renaissance masters and their numerous followers in the Western tradition seemed to acquire a peculiar obsession with depicting architectural elements in their paintings. In this study we proposed and tested the hypothesis that architectural settings might have helped to cover up minute perspective errors or inconsistencies on the figural part of the painting. A total of nine paintings, mostly from Renaissance era, were used as stimuli of this study. The figural theme on the foreground of each painting was cut out from the original image. The remaining blank area was carefully filled in by interpolating the background details. Thus two separated pictures, one with the figure only, another with the background and devoid of the figure, were obtained. A method of limit procedure was used to measure the viewer's sensitivity to detecting perspective errors. The figural theme were either presented on a blank or on the original background. We found that the tolerance for errors was lower in the blank background condition. Our results indicate that imposing architectures might evoke a dominant impression of perspective-correctness, which effectively masks the detection of perspective errors on other parts of the painting.

What 80 Lisas can reveal about Leonardo's Mona Lisa: One step further in demystifying La Gioconda's absorbing smile

Vera M. Hesslinger, Rüdiger Görlitz and Claus-Christian Carbon

From an art historian perspective, Gombrich has already unraveled the main characteristic of Mona Lisa's smile: Ingeniously using local sfumato technique, that is creating primarily low spatial frequencies around the corners of the mouth, Leonardo da Vinci created the impression of a smile emerging only when the mouth is perceived peripherally. Until now, only an eyetracking experiment tested Gombrich's hypothesis by applying eye-contingent display changes to simulate the dissociate perception of the expression of the mouth when perceived peripherally (smile) versus centrally (neutral) [Bohrn et al, 2010, Psychological Science, 21(3), 378-380]. The present study follows a more direct approach by generating out of a sample of 20 female faces 2 [NeutralMouth vs SmilingMouth] x 2 [NeutralSfumato vs SmilingSfumato] = 80 versions, with the NeutralMouth-SmilingSfumato-version ('MonaLisa-version') as an analogue to the specific application of sfumato in Leonardo's masterpiece. Participants rated the MonaLisa-version as being the most mysterious of all versions, and as being significantly more smiling and beautiful than the respective NeutralSfumato-version. The specific combination of the perception of intensified smile and beauty together with a strong feeling of mysteriousness induces strong dynamics in the perceptual system, which seems to be the reason for generating the ongoing fascination for La Gioconda.

Simple and constructive visual mental imagery are behaviorally and neurally separable

Alexander Schlegel, Peter Kohler, Sergey Fogelson and Peter Tse

If you ask a bonobo what you get when you cross an elephant with a rhino, he likely won't have the slightest clue. If you ask me, however, I can vividly imagine an elephino with two tusks and two horns, charging through the savanna. Human creative activities such as language, art, and science indicate that we have evolved robust machinery for synthesizing new concepts. But we know little about the neural basis of this machinery. Here we use fMRI and DTI to investigate a specific example of human creativity we call constructive imagery: the synthesis of novel mental visual images given only their component parts. We developed a set of segment stimuli that could be assembled in 2x2 arrays to generate complete figures ranging from simple to complex. Subjects performed two tasks with these stimuli: they either viewed sequentially-presented complete figures and held them in working memory, or mentally constructed a figure during sequential viewing of its four segments. Subjects were required to choose the previously seen or constructed figure from among distractors after a delay. Performance in the two tasks was not significantly correlated, and we identified several brain regions that were selectively recruited during constructive imagery.

There's more than one way to irritation! An attempt to categorize ambiguity in art

Claudia Muth and Claus-Christian Carbon

Irritation is not definable by outside-characteristics but is elicited by inconsistencies between internal states [Zschocke, 2006, Munich, Fink]. The principle of exclusivity calls for only one clear interpretation at a time thus explaining why we perceive a sudden switch between interpretations when bi-stable pictures are perceived. Zeki [2004, Consciousness and Cognition, 13(1), 173-196] defines ambiguity accordingly as the existence of multiple possibilities or schemata in contrast to uncertainty. We extracted several categories of ambiguity out of multi-dimensional semantic scales gathered for the perception of modern art that go beyond ambiguity as a simple 'switch' of interpretations. Examples out of experiential reports further indicate that contradictory elements can be found on several levels of processing including imagination, association and simulation processes. We propose that modern art is capable of fusing unsolvable contradictions in various ways. Several theories suggest that elaboration of these ambiguities [Ramachandran and Hirstein, 1999, Journal of Consciousness Studies, 6(6-7), 15-51] and specifically the reduction of prediction-errors [Van de Cruys and Wagemans, 2011, i-Perception, 2(9), 1035-1062] are linked to aesthetic appreciation which underlines the relevance of looking closer at the several ways of irritation in art.

How heavy is red? Towards a psychophysical framework of color weights

Michael Klaus Groh and Claus-Christian Carbon

From a physical point of view, colors derive from the spectrum of light and are by definition zero-g. As vision scientists are quite uninterested in definitions of mere physical properties but try to understand and analyze the empirical world caused and framed by the cogni-perceptual apparatus, we aimed to measure the weight of certain colors. To enable detecting even subtle differences between 'psychological weights' of colors we decided to use relative weight judgments of two identical spheroids per trial. Physical weight was re-measured after each experimental session with a precision of <0.1 g for targets of 70 g. Furthermore, participants were checked for personality traits and their affective state before and after the testing. Participants showed highly consistent relative weight estimations, but only if they had low scores on the neuroticism scale: then they estimated violet and black as heaviest, followed by blue, white, green and orange with the lightest colors being red and yellow. This data pattern is quite remarkable as weight estimation was not directly related to important visual color aspects saturation or lightness, but strongly related to energy of light measured in electronvolt (eV) opening up the possibility for establishing a psychophysics of color weights.

Naive observers' perception of beauty, glossiness and inference color of pearls: Comparison with expert appraiser

Michiteru Kitazaki, Kaori Yanase, Yusuke Tani, Takehiro Nagai, Kowa Koida and Shigeki Nakauchi

Aesthetic perception is influenced by processing fluency (Reber et al, 2004, Personality and Social Psychology Review, 8(4), 364-382). For instance, symmetric figures are preferred and judged beautiful. We focused on aesthetics perception of real objects such as pearls. Experts appraise pearls' beauty. However, it has not been investigated whether the appraisement reflects naive observers' aesthetic perception. We aimed to investigate how naive observers judge beauty of pearls in relation to perception of glossiness and inference color, and whether it correlates with expert's appraisement. Twenty-three naive participants ranked nine pearls based on beauty, then ranked them based on perceived glossiness and inference color. They were unaware of glossiness and color judgments until completing the beauty judgment. Judged beauty correlated with image skewness (glossiness index) and inference-color index (partial correlation, skewness: r=.36, inference color: r=.25), but it was weaker than expert's ranked beauty (skewness: r=.78, inference color: r=.90). Perceived glossiness well correlated with both skewness (r=.62) and inference color (r=65). Thus, expert's appraisement more correlated with perceived glossiness (r=.77) than judged beauty (r=.13) and perceived inference color (r=-.17). These results suggest that our naive perception of glossiness would be a source of experts' appraisement since glossy objects are salient and processed fluently.

Effects of mere exposure and abstractness on aesthetic preference of visual stimuli in children

Dragan Jankovic

According to the two-step theory of the mere exposure effect, repeated exposure to a stimulus enhances subjective feeling of perceptual fluency, which in turn influences preference for old over new stimuli. However, the findings obtained in children are not so ubiquitous. The present study was aimed at examining relations between two different perceptual fluency measures and aesthetic preference of visual stimuli in 9 year old children. Perceptual fluency was manipulated in two ways, by abstractness (visual stimuli with high and low content accessibility) and frequency of exposure (0, 2 and 5 exposures) in within-subjects design. Following the exposure phase, participants judged presented stimuli on the seven-step bipolar beautiful-ugly scale. The results showed that effects of mere exposure were opposite for stimuli with high and low content accessibility. Repeated exposure to a stimulus with high content accessibility increases the liking, while repeated exposure to a stimulus with low content accessibility decreases the liking.

Perceptual Fluency does not necessarily increase aesthetic appreciation: Evidence against the Hedonic Fluency Model

Sabine Albrecht and Claus Christian Carbon

Aesthetic pleasure is a function of the perceiver's processing dynamics. According to the Hedonic Fluency Model [Winkielman et al, 2003, in: The psychology of evaluation: Affective processes in cognition and emotion, J Musch and K C Klauer, Mahwah, Erlbaum], the perceptual fluency of a stimulus has an influence on affective judgments about this stimulus. This so-called Fluency Affect Link predicts higher positive judgments with increasing perceptual fluency. In the present experiment, stimuli of different levels of complexity and of positive as well as negative valence from the IAPS database were tested for their aesthetic appreciation. Perceptual fluency was manipulated through perceptual priming. Contrary to the predictions of the Hedonic Fluency Model, participants'(n=20) aesthetic appreciation only benefited from fluency of stimuli with positive, but not with negative valence. Actually, stimuli with negative valence were judged more negative in the highly fluent than in the low fluent condition. These findings are compatible with the Fluency Attribution Hypothesis [Jacoby et al, 1989, in: Varieties of memory and consciousness: Essays in honour of Endel Tulving, H L Roediger and F I M Craik, Hillsdale, Erlbaum], asking for an adaptation of current research hypotheses often derived from the Hedonic Fluency Model.

Systematic assessment of emotional expression using facial animation techniques

Kenneth C. Scott-Brown, Brian Robinson, Moore Fhionna, Malcolm Cook, Santiago Martinez and Robin J. S. Sloan

Research on emotional expression tends to use static faces and classic emotions. We report a series of experiments using digital animation techniques to assess the extent of emotional expression recognition for a range of artistically defined emotional expression choreographies varying the sequence of animation from the eyes and the mouth. Artistic theory prediction suggests that authentic emotional expression should be led by the upper half of the face. We constructed dynamic emotional expressions with 5 facial animation sequences, eyes then mouth, eyes and mouth overlapping in time, simultaneous mouth and eye animation, overlapping mouth and eyes or separate mouth and eyes. The six emotional expressions were presented in a random sequence to 30 observers. Emotional expressions were recognised above chance, and authenticity judgements favoured upper face leading animations. Given a secondary emotion label opportunity, disgust and fear were included in each-others descriptions, consistent with artistic prediction. The results show how a systematic approach to facial animation can test predictions from the artistic method. In addition, the use of computer-generated faces can be used to create emotional test stimuli that are able to test the extent of emotional expression recognition in a systematic manner not possible with human actors.

The influence of a product's perceived social function on aesthetic pleasure for visual product designs

Janneke Blijlevens and Paul P.M. Hekkert

Aesthetic pleasure for product designs is often researched through manipulating visual properties of design (e.g., symmetry and complexity). We contribute by researching how perception of a product design's social function influences aesthetic pleasure. People have the inherent social needs to feel autonomous and to feel connected to others (Brewer, 1991, Personality and Social Psychology Bulletin, 17, 475-482) and use product designs to express their autonomy or to whom they are connected (Belk, 1988, Journal of Consumer Research, 15, 139-168.). From an evolutionary perspective, product designs provide aesthetic pleasure, because it helps direct beneficial behavior for people's survival (Tooby and Cosmides, 2001, Substance, 94/95, 6-27). Consequently, we argue and investigate that product designs that help fulfill people's social needs are aesthetically pleasurable. Respondents rated visual stimuli of either headphones or sunglasses on the level to which they helped them feel autonomous and connected and on how aesthetically pleasurable they were. Results show that both a product's level to express someone's autonomy and level to express connectedness positively influence aesthetic pleasure for visual product designs. Concluding, a visual product design is aesthetically pleasurable when people perceive that a product design may help in fulfilling their inherent social needs.

Errors in visual search lead to devaluation of stimuli

Andrey Chetverikov

People automatically monitor and evaluate their performance in visual tasks, and this process is closely associated with negative affect [Luu et al, 2000, Journal of Experimental Psychology, 129, 43-60]. I tested the hypothesis that this meta-cognitive monitoring will influence the preference ratings for target and distractor stimuli. In the first experiment participants located target face in a circular array of faces presented for 600 or 1200 ms. Targets were specified by sex and tint (blue or yellow). After the answer participants evaluated target and distractors. The results indicated that 1) distractors were evaluated less positive than targets, and 2) stimuli were rated worse in case of incorrect answers. In the second experiment I used more complex stimuli. The miniature paintings were presented for 1600 or 2200 ms; target was specified by type (portrait or landscape) and by plus or minus symbol. Only the effect of answer correctness was replicated, but it was moderated by participants' overall accuracy. It is argued that as the exposure time was fixed, accurate participants have had to make additional effort to concentrate on target after it has been located. This, in turn, leads to target devaluation and the observed interaction. Supported by RFBR grant #11-06-00287a.

Six tea-towels, one calendar and 1659 children's self-portraits: A developmental study of children drawing faces

Chris McManus, Rebecca Chamberlain, Charlotte Christopherson, Lucy Prince-Mobbs, Mark Robinson and Ines Stelk

Children and adults typically do not draw faces accurately, the commonest error being that the eyes are placed too close to the top of the head, instead of veridically at about halfway down the head. This study looked at 1659 full-frontal self-portraits which were drawn by children aged 3 to 10 in a London primary school during the autumn of 2007-10 and printed in a calendar or on tea-towels and sold for fund-raising at Christmas by the Parent-Teacher Association. Additionally, 81 self-portraits drawn by adult teachers were also available. The positions of 24 landmarks from eyes, ears, nose, mouth and head outline were digitised from photographs and used for statistical analyses. Comparative, normative data from real heads were obtained from the Farkas and Munro anthropometric data used in paediatrics. With increasing age, the drawn heads became more oval, and eyes and nose were placed lower, although the mouth stayed at the same relative position. Eyes also became wider and more oval, the corners of the mouth turned up less, and the ears were placed higher in the head. Although drawings became closer to the proportions of actual faces, even in adults the eyes continued to be placed too high.

Viewing distance and angle matter in the Poggendorff illusion

Olga Daneyko, Daniele Zavagno and Natale Stucchi

In a previous study (Daneyko et al, 2011, i-Perception, 2, 503-507) we argued that the geometrical misalignment of two parts of the long cross in the Byzantine mosaic known as Lunette of San Lorenzo (Ravenna) is related to the Poggendorff illusion, which induced the artist to prefer a perceptual adjustment over a geometrical one. However, the geometrical misalignment was greater in the original mosaic than the average alignment determined by participants to our previous experiment. In the present work we addressed this difference by projecting a silhouette rendering of the mosaic on a large screen, simulating both size and height of the saint viewed from ground level. Participants were asked to align one end of the cross to the other end from two different viewpoints: 1) from a distance of 50 cm frontoparallel to the projection, and 2) from a distance of 750 cm, 450 cm below the projection. Results from condition 1 were similar to those obtained in our previous experiment, while results from condition 2 are consistent with the geometrical misalignment of the original mosaic. Our findings support the hypothesis that the artist was aware of the misalignment illusion, noticed when he inspected the mosaic from ground level.

The Inner Scribe: A role for visual short-term memory in observational drawing?

Rebecca Chamberlain, Chris McManus, Howard Riley, Qona Rankin and Nicola Brunswick

Enhanced visual perception is critical for successful observational drawing (Kozbelt, 2001, Visual Cognition, 8, 705-723). However, the importance of visual memory (VM) in observational drawing remains under dispute. Empirical evidence suggests that there is a correlation between performance on incidental long-term visual (VLTM) memory recall and drawing accuracy (McManus et al, 2010, Psychology of Aesthetics, Creativity and the Arts (PACA), 4, 18-30). Conflicting evidence suggests that maintenance of visual information during drawing is circumvented by a more direct perceptual system, as artificially lowering gaze frequency affects drawing accuracy (Cohen, 2005, Perception and Psychophysics, 67, 997-1009). The former link between VLTM and drawing ability may be explained by the role of VLTM in online scene perception and enhanced perceptual analysis (Hollingworth, 2004, Journal of Experimental Psychology: Human Perception and Performance, 30, 519-537) rather than in the ongoing maintenance of visual stimuli, thereby resolving this conflict. However, evidence of a connection between visual short-term memory precision (VSTM) and observational drawing could not be explained in the same manner and would be suggestive of a role for VM above and beyond perceptual guidance. Short-term reproduction and delayed match-to-sample paradigms using simple geometric shapes and line pairs were used to test this contention. The results of these studies are discussed in relation to recent findings that reveal an encoding advantage for expert artists (Glazek, 2012, PACA, online first).

Evaluation of Graffiti/Art in different Contexts

Andreas Gartus and Helmut Leder

According to the model of Leder et al. [2004, British Journal of Psychology, 95, 489-508], the pre-classification of an image as an object of aesthetic interest is a necessary pre-condition of the aesthetic process. This requires adequate context variables like e.g. the appearance in a museum or an art gallery. In this work we investigate the influence of the presentational context on some aspects of the process of aesthetic appreciation. We presented graffiti / street art and more established artworks to our participants - either integrated in a street or in a museum scene. The participants were asked to rate these pictures (without scene context) on the three 7-point scales 'liking', 'interest' and negative to positive 'emotion'. We found a significant interaction of context type and interest in graffiti (assessed by a questionnaire prior to the experiment) for the third scale ('emotion'). That is, participants with high interest in graffiti reported a more positive emotion than participants with low interest in graffiti when the stimuli were presented in the street context (vs. the museum context). These results suggest that different presentational contexts can indeed have different effects depending on individual attitudes towards the stimulus types.