Implementation of Human Tracking System under Different Video Degradations

Abstract

In this paper, a new strategy for human tracking is proposed to enhance the human detection and tracking in the presence of video degradations using Kalman filter. The average number of assigned tracks (indicating the correct detection and tracking) was calculated. We have measured the human tracking accuracy in the presence of both speckle and salt & pepper video disturbances. Experimental results show that both median and disk filters have targeted the most efficient masking system that easily restore the accurate number of the average detections and predictions.