Real Time Face Detection and Recognition

Security and surveillance are the two important aspects of human being in day to day life. In this project, it has been proposed that face detection and recognition system will capable of processing images very fast at 15 or 20 frames/sec with 544 x 288, 640 x 480 or any resolution from C270HD Logitech webcam, while acquiring maximum true detection rate and very low false positive rate. In our implementation system, we followed three aspects which are important. First one is Viola and Jones algorithm for face detection, second is principal component analysis algorithm for facial feature extraction purpose and last is data base created by us. We have taken total of 80 images in database with pixel size (resolution) of 182 x 182. Total of 8 subjects each having 10 images with different aspects like illumination, head angles, obstacles etc. Euclidean distance measurement technique is used for excellent matching between input image and database image. Minimum is the distance maximum is the matching. System has overall classification accuracy of 87.5% with false acceptance of 91.6667% and false rejection of 87.5%.