Abstract: In this paper, the target is considered as human, we propose a multiframe marked shape model based human detection and tracking. Geometric features are measured through principle component analysis features are reduced to half of its ratio and finally a realtime decision making system detects human and by representing region of interest through bounding box followed by tracking in order to speed up the process. Optical tracking method is used for tracking purpose. To shrink or expand the bounding box as required, the aforementioned sample rectangles generated during the tracking process and their associated confidence ratings which indicate the likelihood of a rectangle containing the region of interest Finally our proposed method, shape based target detection algorithm successfully tracking a human target walking across the view of the camera.
Keywords: human detection, tracking, realtime, shape characteristic recognition.
Title: Target Detection and Tracking in Real Time with Multi frame Marked Point Process Model
Author: N. Mohanambigai, M. Arul pugazhendhi
International Journal of Electrical and Electronics Research
ISSN 2348-6988 (online)
Research Publish Journals
