ICTACT Journal on Image and Video Processing

A. Thamizharasi ¹, J. S. Jayasudha ²

Affiliations
1 Department of Computer Science and Engineering, Manonmaniam Sundaranar University, IN
2 Department of Computer Science and Engineering, Sree Chitra Thirunal College of Engineering, IN

Abstract

Face recognition system is gaining more importance in social networks and surveillance. The face recognition task is complex due to the variations in illumination, expression, occlusion, aging and pose. The illumination variations in image are due to changes in lighting conditions, poor illumination, low contrast or increased brightness. The variations in illumination adversely affect the quality of image and recognition accuracy. The illumination variations in face image have to be pre-processed prior to face recognition. The Contrast Limited Adaptive Histogram Equalization (CLAHE) is an image enhancement technique popular in enhancing medical images. The proposed work is to create illumination invariant face recognition system by enhancing Contrast Limited Adaptive Histogram Equalization technique. This method is termed as "Enhanced CLAHE". The efficiency of Enhanced CLAHE is tested using Fuzzy K Nearest Neighbour classifier and fisher face subspace projection method. The face recognition accuracy percentage rate, Equal Error Rate and False Acceptance Rate at 1% are calculated. The performance of CLAHE and Enhanced CLAHE methods is compared. The efficiency of the Enhanced CLAHE method is tested with three public face databases AR, Yale and ORL. The Enhanced CLAHE has very high recognition accuracy percentage rate when compared to CLAHE.

Keywords

Illumination Invariant, Face Recognition, Contrast Limited Adaptive Histogram Equalization (CLAHE), Enhanced CLAHE, Fisher Face.

Full Text

     

A. Thamizharasi ¹, J. S. Jayasudha ²

Affiliations
1 Department of Computer Science and Engineering, Mohandas College of Engineering & Technology, IN
2 Department of Computer Science and Engineering, SCT College of Engineering, IN

Abstract

The face recognition is popular in video surveillance, social networks and criminal identifications nowadays. The performance of face recognition would be affected by variations in illumination, pose, aging and partial occlusion of face by Wearing Hats, scarves and glasses etc. The illumination variations are still the challenging problem in face recognition. The aim is to compare the various illumination normalization techniques. The illumination normalization techniques include: Log transformations, Power Law transformations, Histogram equalization, Adaptive histogram equalization, Contrast stretching, Retinex, Multi scale Retinex, Difference of Gaussian, DCT, DCT Normalization, DWT, Gradient face, Self Quotient, Multi scale Self Quotient and Homomorphic filter. The proposed work consists of three steps. First step is to preprocess the face image with the above illumination normalization techniques; second step is to create the train and test database from the preprocessed face images and third step is to recognize the face images using Fuzzy K nearest neighbor classifier. The face recognition accuracy of all preprocessing techniques is compared using the AR face database of color images.

Keywords

Illumination Normalization, Contrast Stretching, Power Law, Homomorphic Filter, Log Transformations, FKNN Classifier.

Full Text