Caitlin Mullin

In the Eye of the Beholder: Rapid Visual Perception of Real-Life Scenes by Young Adults with and Without ASD

Typically developing (TD) adults are able to extract global information from natural images and to categorize them within a single glance. This study aimed at extending these findings to individuals with autism spectrum disorder (ASD) using a free description open-encoding paradigm. Participants were asked to freely describe what they saw when looking at briefly presented real-life photographs. Our results show subtle but consistent group-level differences. More specifically, individuals with ASD spontaneously reported the presence of people in the display less frequently than TD participants, and they grasped the gist of the scene less well. These findings argue for a less efficient rapid feedforward processing of global semantic aspects and a less spontaneous interpretation of socially salient information in ASD.

There is beauty in gist: An investigation of aesthetic perception in rapidly presented scenes

While an artfully crafted painting can evoke profound aesthetic experience, the same applies to a grand ballroom or sunset. Like fine art, everyday scenes contain aesthetic qualities, with some scenes preferred over others. The meaning or semantic label of a scene, known as scene gist, is extracted rapidly and automatically, with just a brief glance, computed mainly using low spatial frequencies (LSF) in the image. Although we can easily identify a scene, the question remains, if an accurate aesthetic impression can be formed from such rapid and coarse overall representation. We investigated the characteristics of scene gist to determine if aesthetic preference can be extracted with such short display durations. Furthermore, given that scene gist is based on an initial coarse representation, we asked whether LSF renderings of these scenes would elicit similar aesthetic judgments. Using a between-groups design, we found a significant positive correlation between aesthetic judgments on real-world scenes for images viewed for an unlimited amount of time and those viewed for only 45ms, but no significant correlation with the LSF set. This demonstrates that aesthetic judgments can be extracted rapidly and are relatively stable across display durations but do not survive image degradation, suggesting that image content outweighs structure. We performed the Implicit Associations Test by using aesthetically pleasing and non-pleasing images from the previous experiment paired with aesthetically pleasing and non-pleasing words, to examine whether these aesthetic judgments are also made automatically when they are irrelevant to the task. Participants made significantly more classification errors and were slower when pleasing scenes were paired with non-pleasing words. This suggests that participants could not help but make aesthetic judgments on real-world scenes. Additionally, we found that the most pleasing and non-pleasing scenes differed significantly on self-similarity and anisotropy, measures of image statistics relating to computational aesthetics. Meeting abstract presented at VSS 2015.

Gist Perception of Image Composition in Abstract Artworks

Most recent studies in experimental aesthetics have focused on the cognitive processing of visual artworks. In contrast, the perception of formal compositional features of artworks has been studied less extensively. Here, we investigated whether fast and automatic processing of artistic image composition can lead to a stable and consistent aesthetic evaluation when cognitive processing is minimized or absent. To this aim, we compared aesthetic ratings on abstract artworks and their shuffled counterparts in a gist experiment. Results show that exposure times as short as 50 ms suffice for the participants to reach a stable and consistent rating on how ordered and harmonious the abstract stimuli were. Moreover, the rating scores for the 50 ms exposure time exhibited similar dependencies on image type and self-similarity and a similar pattern of correlations between different rating terms, as the rating scores for the long exposure time (3,000 ms). Ratings were less consistent for the term interesting and inconsistent for the term pleasing. Our results are compatible with a model of aesthetic experience, in which the early perceptual processing of the formal aspects of visual artworks can lead to a consistent aesthetic judgment, even if there is no cognitive contribution to this judgment.

The artistic Turing test: An exploration of perceptions of computer-generated and man-made art

The generation of genuinely creative works of art could be considered as the final frontier in artificial intelligence (AI). Several AI research groups are pursuing this by programming algorithms which generate works in various media and styles. The ultimate test of success for such a machine artist would be to convince the onlooker that it was generated by a human being; an artistic Turing test. Previous research has shown that observers can distinguish artworks generated by a skilled human artist over those of a child or an animal using the perception of intentionality- the appearance of a planned final product (Hawley-Dolan & Winner, 2011). However, there is a little research on whether observers can differentiate between computer-generated art and man-made art, and if so, whether these judgments are driven by impressions of intentionality or surface characteristics that identify the mode of production. Furthermore, the decision that an artwork is computer-generated may reflect a negative aesthetic preference, as research has suggested that believing an artwork or musical composition to be computer-generated negatively affects its aesthetic appraisal (Kirk et al, 2009; Moffat & Kelly, 2006). The current study examined whether individuals were able to differentiate between works of art whose creative or representational abilities are defined by computer algorithms, from matched artworks created by human artists. Participants sorted artworks into computer- or human-generated and indicated their aesthetic preference on a 7 point Likert-scale. Results show that participants were able to successfully determine the provenance of the artworks. Perception of intentionality and surface characteristics as well as a subset of image statistics (Pyramid of Histograms of Orientation Gradients (PHOG), luminance spectra) were also investigated in relation to source decision criteria. The results have implications for the way in which AI algorithmic art is created, as well as providing insights into their aesthetic perception. Meeting abstract presented at VSS 2015.