Facial image database mining and classification analysis using different distance metrics

This paper discusses two important works. First work deals with extraction of key features from standard face databases. The data mined from the face databases consist of information related to number of facial images, male and female, different facial expressions, different poses, illumination conditions, facial image file format, synthetic faces, masked faces, infrared faces and number of subjects. The second work involves testing the recognition rate of faces available in these databases. It compares the recognition accuracies obtained by different distance metrics such as L1 norm, L2 norm and Mahalanobis distance. Along with classification rate, the other parameters such as mean squared error, peak signal to noise ratio and rank are also compared for different face databases. Tested results are tabulated and plotted and a detail analyzed report is given in this paper.