Inferotemporal network model for 3D object recognition

One of the essential cognitive abilities of the human brain is the ability to recognize 3D objects. However, the exact neural mechanism is still unclear. In this study, a model of the inferior temporal cortex for object recognition is proposed. The architecture of the model is based on the modular network SOM. The model has been trained to perform object recognition and classification tasks that have been reported to occur in the TE area in the brain of macaque monkeys. We show by simulation that the model is not only capable of classifying artificial 3D objects according to their structural similarity, but also capable of achieving view-invariant recognition of the objects by incorporating a mental-rotation type mechanism.