Perceptual Scale

Perceptual scaling aims to quantify how humans perceive differences between stimuli, mapping physical properties to subjective experience. Current research focuses on refining methods like Maximum Likelihood Difference Scaling (MLDS) to improve accuracy and comparability across diverse stimuli, and on exploring the relationship between perceptual scales and the internal representations learned by deep neural networks, such as ResNets and Vision Transformers. These investigations are crucial for advancing our understanding of human perception and have implications for improving image processing, computer vision, and the design of more human-centered technologies.