Psychology 2135A/B Lecture Notes - Lecture 5: 2.5D, Retina, Two-Streams Hypothesis

• School: Western University
• Department: Psychology
• Course: Psychology 2135A/B
• Professor: John Paul Minda

Document Summary

Higher-level vision and object recognition low level vision: what happens computationally high level vision: requires objects, lets you know about knowledge. Recovering the third dimension computational processes for reconstructing a 3d world from a 2d retinal image requires several assumptions about the world. Objects are lit from above looks different than normal when things are lit from below we assume things are light from above naturally we assume things are light from above naturally. Objects are attached there are not a lot of natural free floating objects you assume the block is sitting on top, not going through. Perceptual constancy combines the knowledge and the retinal image objects further away look smaller but we assume they are the same size. High level vision bottom up processes: data driven, depends on visual perception bottom up processes: data driven, depends on visual perception top down processes: conceptually driven, depends on knowledge.