Independent Component Analysis (ICA) Based Face Recognition System

Advancement in computing and telecommunications technologies, digital images and video are playing vital role in the present information era. Detection of a person from an image of their face is termed as face recognition. It is based on the classifiers applied for the feature extraction. It has its applications on various domains. Face recognition application without human intervention distinguish human faces in still images, or from a video frame or file. Identification of faces from the videos becomes more complex since the video frames are extensively huge in volume. Recognizing and tracking of face objects in video sequences is still an active research, which is considered in this paper. Illumination and pose is the two major problems exist with the face recognition. The video sequence is divided into frames for the purpose. The paper proposed a new method for face recognition, with the aid of Independent Component Analysis. Numerous faces with different texture, color and illuminations were considered for the detection purpose. The ICA algorithm is applied to recognize the face from the video images. The experimental results reveal that the proposed method yields better results than the existing ones.