Beauty and structural complexity
Samy Lakhal, Alexandre Darmon, Jean-Philippe Bouchaud, Michael Benzaquen
Abstract
The authors study the problem of the relation between image appreciation and complexity. A large-scale survey on two sets of random images reveals that maximum appreciation is obtained at intermediate entropic complexity. A coarse-graining procedure, capturing structural complexity while abstracting from high-frequency noise, appears to be a good predictor of preferences, suggesting the existence to some extent of universal quantitative criteria for aesthetic judgement.
Topics & Concepts
Relation (database)Structural complexityBeautyMathematicsComputer scienceArtificial intelligenceImage (mathematics)EpistemologyStructural approachComputational complexity theoryAestheticsKey (lock)Computer visionCharacterization (materials science)Aesthetic Perception and AnalysisVisual Attention and Saliency DetectionGenerative Adversarial Networks and Image Synthesis