A Model of Visual Object Recognition in the Presence of Active 
Eye Movements

Arash Fazl
Department of Cognitive and Neural Systems
Boston University

Many models of 3D object recognition propose collecting and learning 
several 2D views of the object's surface to achieve an invariable 
representation. But in biological visual systems, an important source 
of variability is the result of eye movements on a surface. Our 
modeling work addresses different problems that arise when we take 
into account space-variant cortical representation and active eye 
movements in a visual object recognition model. We suggest how 
different brain areas in the Ventral and Dorsal visual streams act 
together to explore and learn a 3D surface and achieve object constancy.