Accepted Papers

  • Fast Object Tracking Algorithm for an Embedded System
    Mohamed ELBAHRI and Nasreddine TALEB,Djillali Liabes University,Algeria
    Object tracking is a challenging task used in video surveillance, to help human to detect the abandoned luggage, recognize suspicious persons or track people. In this paper, we propose an approach based on the detection and the recognition. For the detection step, background subtraction is used to delineate the areas containing the motion and to reduce the processing time for the next step. To recognize the detected objects, Speed Up Robust Features (SURF) is used as local feature detector to describe and identify any object. These features are used for their robustness to the illumination and scale changes. To make this system more effective, we propose to implement our program on an embedded system which can be used in suspicious places. Experimental results demonstrate the efficiency of the proposed approach based on both the background subtraction and the SURF features for the object tracking on an embedded system