A Proposed Rdit Algorithm for Analysis of Face Verification in Videos in Image Database for Crime Investigation

In the face recognition most researches has considered about face detection and identification or a face detection and tracking, but best of our knowledge very few researches has focused on the face detection, identification and tracking. The combination of these three mechanism are been used in many real time applications .Here we consider one of the application as theft detection in real time which proposes a new mechanism named as FACE RDIT (Face Rectangular Detection, Identification and tracking) algorithms. Also here we include some preprocessing algorithms like gray scale conversion and histogram equalization for image enhancement process. Apart from this we also derive a new algorithm which is based on rectangular features of the face and multi linear training. Also the proposed technique has four main steps like training, detection, identification and tracking by using this proposed method we train the unauthorized users photo and then make the detection of the faces in the group of frames and then identify the particular personality from the group of images. Finally it tracks the identified thief in next following video frames by our proposed algorithm. Thus the effectiveness of the proposed methods is well supported by both detailed analysis and extensive experimentation on a face verification problem.