An Efficient Face Detection and Recognition Method for Surveillance
K. V. Arya, Abhinav Adarsh · 2015
In this paper a method is presented for automatic detection and recognition of human faces for surveillance purpose. The proposed method first detects skin regions in the image using a skin color model using YCbCr and HSV color space. Then apply height to width ratio followed by face region identification. Lastly PCA verification algorithm is used to detect face accurately. Train face images are used to generate feature space (face space). Test images are then projected on sub spaces and distances measured to find out best match from train images. The face space is affine subspace and face images can be represented as weighted sum of these sub spaces. The proposed method provides ability to detect, extract and recognize face from images taken by camera or video automatically. Recognition under different environmental conditions can be achieved by training on limited number of characteristics faces. The proposed approach is simple, efficient, and accurate as well as relatively insensitive to changes in the face images.