An integrated robust approach for fast face tracking in noisy real-world videos with visual constraints

Face detection and tracking algorithms are used in computer vision applications due to the fact that they provide reliable and fast results. This paper describes a model for face tracking in video sequences using Open Source Computer Vision (OpenCV) software library. To increase the face tracking accuracy, we propose a real time face tracking algorithm based on integration of Continuously Adaptive Mean Shift (CAMShift) and kalman filter. First, haar cascade detects face in the video sequence; once the face is detected then other parts like eyes, nose and mouth are detected. After successful face region detection, key-points are extracted using Speeded Up Robust Features (SURF) framework and CAMShift algorithm applied. CAMShift algorithm calculates width, height and (x, y) coordinates of the face region and this information is provided to kalman filter. Kalman filter calculates new center and size of the bounding box. Based on the center information, face tracking takes place in further frames. The experimental results at the end of this paper clearly indicate that the proposed algorithm integrating CAMShift and kalman filter is better to each single approach. Using this algorithm, we can achieve faster speed in face tracking.