A LITERATURE SURVEY ON VARIOUS ILLUMINATION NORMALIZATION TECHNIQUES FOR FACE RECOGNITION WITH FUZZY K NEAREST NEIGHBOUR CLASSIFIER

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.

Authors
A. Thamizharasi1, J.S. Jayasudha2
Mohandas College of Engineering & Technology, India1, SCT College of Engineering, India2

Keywords
Illumination Normalization, Contrast Stretching, Power Law, Homomorphic Filter, Log Transformations, FKNN Classifier
Published By :
ICTACT
Published In :
ICTACT Journal on Image and Video Processing
( Volume: 5 , Issue: 4 )
Date of Publication :
May 2015

Creative Commons License
This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.