Rotation invariant face recognition and application to facial expressions using optical neural networks

Abstract

Face and facial expression recognition has been an attractive field newlineof research for both neuroscientists and computer vision scientists The newlineautomatic face recognition is the important one In real world the faces to be newlineidentified may be in rotated form So a unique approach for identifying newlineupright and rotated faces is necessary Neural networks in combination with newlineaffine transform can be effectively used for rotating the faces Also neural newlinenetwork has the powerful capability of adaptation The degree of rotation newlinerequired is determined by the neural networks newline newline