A Novel Approach for Recognizing Facial Structures Using Extreme Learning Machine Algorithm

A facial recognition system is a computer application that deals with identification or verification of a person from digital image. To eliminate the error condition, the application of face recognition system is widely used. The altering facial appearance is one of the main challenging problems faced in face recognition algorithms. The proposed system consists of three stages namely; face detection, feature extraction and classification providing face recognition. Initially, the face is detected from the input image and the architectural features are extracted from the face detected image. Using classifier, the original image and morphed images is classified in reference with the features extracted from the detected face image. Finally, the system performs the retrieval of the original face from the morphed face. The proposed extreme learning machine algorithm yields high accuracy in face recognition technique as compared to the other algorithms. The proposed algorithm yields proper high detection and recognition accuracy as compared to different existing face algorithms.