Modeling Holistic Object Recognition
Natural object recognition is very hard to modulate and understand because of its complexity. A robust recognition system requires both selectivity among different object categories and tolerance to vastly different retinal images of the same object, resulting from natural variation in (e.g.) position, size, illumination, and clutter. In this talk, I will, demonstrate the principles behind a holistic object recognition system. Moreover, I shall also give an introduction to which brain functions that are involved in visual recognition – the ventral path. Finally, I will also discuss how language may influence the differentiation process of the categories we are practicing in natural talk.