Description du projet
Apporter le chaînon manquant entre esthétique objective et subjective
La perception de la beauté est quelque chose de très personnel. C’est pour cette raison que cette notion n’a pas fait l’objet de suffisamment de recherches. Il existe deux approches principales pour appréhender l’esthétique. L’une concerne les propriétés statistiques qui sont supposées être universelles et déterminées biologiquement. L’autre fait appel à des concepts tels que le style, la signification et les associations personnelles qui dépendent des influences culturelles, de l’expertise artistique et des expériences individuelles. Le projet GRAPPA, financé par l’UE, appliquera la psychologie de la Gestalt pour développer et tester une hypothèse qui fait le lien entre ces deux approches. L’apprentissage automatique permettra d’analyser de nouvelles données provenant d’études en ligne faisant intervenir de nombreux échantillons d’images et de participants. Cela aboutira à un modèle capable de prédire les préférences esthétiques.
Objectif
"""De gustibus et coloribus non disputandum est."" With this slogan philosophers and lay people alike have dismissed all attempts to understand taste, color perception, or aesthetic preferences. Sense of beauty may just be too individual and too complex to qualify as target of scientific inquiry. Yet, since Fechner (1876), empirical aesthetics has studied the factors determining people’s aesthetic responses to art works and objects, scenes or events encountered in everyday life. Most accounts focus either on high-level concepts such as style, meaning and personal associations, or on low-level statistical properties. While the latter are supposed to be universal and biologically determined, the former are subject to cultural influences, art expertise and individual experiences. Progress in this tradition has reached its limits, which this project overcomes by investigating how Gestalts Relate Aesthetic Preferences to Perceptual Analysis (GRAPPA). Its pioneering working hypothesis is that the way perceivers organize their sensory inputs into meaningful entities (Gestalts) provides the missing link between the two traditional sets of explanations. This hypothesis is fleshed out and tested in a coherent research program linking aesthetic preferences for images of paintings and everyday photographs to general principles of perceptual organization as well as to specific aesthetic concepts like composition, balance and visual rightness. New data from online studies with large samples of images and participants will be analyzed with state-of-the-art computational methods (machine learning) to reveal the critical mid-level factors. This will yield a model to predict aesthetic preference, which will be tested in well-controlled psychophysical and behavioral experiments (e.g. eye-movement recording) and validated also in ecologically richer settings (e.g. in galleries and art museums) and in unconventional cross-over collaborations with contemporary artists."
Champ scientifique
Programme(s)
- HORIZON.1.1 - European Research Council (ERC) Main Programme
Thème(s)
Régime de financement
HORIZON-ERC - HORIZON ERC GrantsInstitution d’accueil
3000 Leuven
Belgique