Abstract:
We propose a neural network of ventral visual pathway for recognition and classification of 3D realworld objects. The model consists of hierarchy of modules, representing the visual areas V1–V4 and inferotemporal cortex (IT). The architecture of the model represents the self-organized map (SOM) of functional radial-basis function (RBF) modules. We show that the proposed architecture is capable of recognizing and categorizing 3D real-world objects in a neurophysiologically plausible way.