Adaptive automatic tracking, learning and detection of any real time object in the video stream

Surveillance is important in most of the sensitive areas for crime detection. Tracking, learning and detection of any real time object in the video stream is an essential part of surveillance. Researchers have designed the object tracking systems for better efficiency but these systems sometime fails due to loss of information caused by complex shapes, rapid motion, illumination changes, scaling and projection of 3D world on 2D image. Apart from these system does not perform well for objects such as foot travelers. This paper proposes a modified PN learning algorithm for enhancing the performance of system. In this algorithm, object is to be tracked as P-Type object and background is divided into a numbers of equal size N-Type objects. It achieves performance improvement by reducing the detector errors. Proposed modified PN learning algorithms achieves performance to increase the frame processing by adding background subtraction technique for any real time object detection. Special feature used in this techniques is Region of Interest (ROI) which minimizes the searching region. Proposed system compares the result using benchmark data set. ROI and background subtraction technique used in this approach minimizes delay for frame processing, increases number of successfully tracked frames, matching percentage and learning index value.