AUTOMATED FACE DETECTION AND TRACKING FACIAL FEATURES IN VIDEOS

In our paper we propose efficient face detection and tracking system that utilizes facial landmark features as key features of the face implemented in KLT algorithm. Sequences Video Frames gives us more information than a single image. Automated detection and tracking of a face using key feature extractors (points) play a significant role in vast variety of applications including Human action recognition, Automated Surveillance system, and human computer interaction. In our methodology, we have developed simple face detection and tracking system by splitting the tracking method in to three parts. 1) Detecting a face, 2) identify facial features for tracking, 3) Track the face. Detecting the face is implemented using Viola Jones object detector, and tracking is done by KLT algorithm. Investigated results shows that our system is more robust and efficient in processing the video frames to various kinds of face rotations and occlusions.