Face Identification using Histogram

Abstract

Various experiments or methods can be used for face recognition and detection however two of the main contain an experiment that evaluates the impact of facial landmark localization in the face recognition performance and the second experiment evaluates the impact of extracting the HOG from a regular grid and at multiple scales. We observe the question of feature sets for robust visual object recognition. The Histogram of Oriented Gradients outperform other existing methods like edge and gradient based descriptors. We observe the influence of each stage of the computation on performance, concluding that fine-scale gradients, relatively coarse spatial binning, fine orientation binning and high- quality local contrast normalization in overlapping descriptor patches are all important for good results. Comparative experiments show that though HOG is simple feature descriptor, the proposed HOG feature achieves good results with much lower computational time.