Juno Kim: “Perceiving objects, their surface reflectance properties and organisation in the environment”
CiNet B1F セミナールーム
We readily perceive the physical properties of objects and their organisation from the pattern of light their surfaces reflect to form images (e.g., 3D shape, colour, gloss and opacity, also occlusion). However, a highly complex inverse problem exists in understanding how we perceptually recover this information from those images. This presentation will introduce some recent evidence to support the view that much of the problem of understanding material appearance can be reduced to a more accessible problem of detecting and classifying image contours. The evidence will show that low-level computations can explain some of this experience, but that we ultimately depend on further mid-level integration of this information to reliably experience the physical properties of objects and their organisation in the environment.