C3 Effective features inspired from Ventral and dorsal stream of visual cortex for view independent face recognition

This paper presents a model for view independent recognition using features biologically inspired from dorsal and ventral stream of visual cortex. The presented model is based on the C3 features inspired from Ventral stream and Itti's visual attention model inspired from the dorsal stream of visual cortex. The C3 features, which are based on the higher layer of the HMAX of the ventral stream of visual cortex, are modified to extract important features from faces in various viewpoints. By Itti's visual attention model, visual attention points are detected from faces in various views of faces. Effective features are extracted from these visual attention points and the view independent C3 effective features (C3EFs) are created from faces. These C3EFs are used to distinct multi-classes of different subjects in various views of faces. The presented model is tested using FERET face datasets with faces in various views of faces, and compared with C2 features (C2SMFs) and C3 features of standard HMAX model (C3SMFs). The results illustrated that our presented view independent face recognition model has high accuracy and speed in comparison with standard model features, and can recognize faces in various views by 97% accuracy.