Face Identification and Verification using Histogram Based Approach in Spatial Domain

Abstract

 The present paper relates to “Face Identification and Verification”. Histogram based image feature extraction approach is used for the system implementation. The database taken for the work comprises of 40×10 size images from forty different objects, that’re limited to 92×112 pixels. Firstly the eyes and face are detected and then cropped for the recognition purpose. The histogram approach performs cumulative distribution to identify the images of eye and face. Therefore a non-linear change that reflects the “accumulative distribution” of the test image produces an image having uniform histogram. Euclidean distance is also calculated for the processing to make the comparison between the test and database image. Some of the difficulties are found in the work excessive changeability in head revolution and slope, intensity and angle of light, facial characteristics, etc. It is observed that the method is best suitable in case of eye recognition.