An improved face recognition method using local binary pattern method

Security system based on biometrics is becoming more popular everyday as a part of safety and security measurement against all kind of crimes. Among several kinds of biometric security systems, face recognition is one of the most popular one. It is one of the most accurate, mostly used recognition methods in modern world. In this paper, two most popular face recognition methods have been discussed and compared using average image on Yale database. To reduce calculation complexity, all training and test images are converted into gray scale images. The whole face recognition process can be divided into two parts face detection and face identification. For face detection part, Viola Jones face detection method has been used out of several face detection methods. After face detection, face is cropped from the actual image to remove the background and the resolution is set as 150x150 pixels. Eigenfaces and fisherfaces methods have been used for face identification part. Average images of subjects have been used as training set to improve the accuracy of identification. Both methods are investigated using MATLAB to find the better performance under average image condition. Accuracy and time consumption has been calculated using MATLAB code on Yale image database. In future, this paper will be helpful for further research on comparison of different face recognition methods using average images on different database.