Abstract
Face recognition is one of the most popular biometric techniques for automatically identifying or verifying a person from a video or digital image. The face recognition accuracy can be affected by intraclass variations and interclass variations. A change in lighting condition is one of the intraclass variations. Preprocessing is an approach to normalize the intraclass variations of light varying image. Histogram equalization (HE) is one of the techniques to normalize the variations in illumination. But it is not suitable for well lighted images. Image quality based adaptive face recognition is used for well lighted face image recognition. The multiresolution property of wavelet transforms is used in face recognition to extract facial feature descriptors. Low and high frequency wavelet subbands are extracted and fusion of match scores from the subband is used to improve the recognition accuracy under varying lighting conditions. For face recognition, 2DPCA (2D Principle Component Analysis) method is used and can be verified with illumination variant face images. 2DPCA is based on 2D image matrices rather than 1D vector so the image matrix does not need to be transformed into a vector prior to feature extraction.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have