PCA Based Improved Face Recognition System.

Due to the rapid evolution of modern image processing and pattern recognition techniques, there exists a variety of biometric techniques like fingerprints, iris (retina) scans, and speech recognition etc. nowadays. However, among them, face recognition is still the most common technique which is in use due to the fact that it is easy to install and has less complexity. It has been a prominent research field for security applications such as video surveillance, fraud detection, person tracking and crowd recognition. This research work discusses and implements the facial recognition by using MATLAB environment in real time. In this work, the face recognition system is implemented using Principal Component Analysis (PCA) and Eigenface approach by dealing with large dataset. A methodology is proposed to produce more accuracy and efficiency by removing the unrelated space from the image. Also, a graphical Interface System (GUI) is developed in order to make our system clearer and to measure the training time. Furthermore, large databases (ORL and private Database named face-100) are tested through PCA and Eigenfaces approach and then the person identification is carried out. The system successfully recognized the human faces and worked better in different conditions like illumination and blur conditions of the face. The rate of male and female accuracy is also calculated.