Real-time robust face detection and tracking using extended haar functions and improved boosting algorithm

A system and method for face detection and tracking is proposed here for the further purpose of facial expression recognition. The method proposed here detects the face in a real time manner from the live video fed through webcam integrated along with the PC. From the live video, a human face is detected first with an improved Haar based algorithm. From the detected face, the areas of interests namely eye brows, eyes, nose and lips are detected. With the help of these features, the method tracks the face and facial components. In the proposed method, the algorithm proposed by Shi and Tomasi is used to extract those features. The method proposed here uses an adaptive version of Lucas-Kanade to track the extracted features. In this paper, the method proposed has performed the face tracking in two steps, first step is the face detection and the second one is face tracking as it will make movements in the live video. The experimental results demonstrate that the proposed algorithm can robustly detect and track human face(s), even in the environments of harsh lighting, unexpected occlusions and vertical or horizontal rotation of face to a considerable extent.