Foreground Object Detection and Tracking for Visual Surveillance System: A Hybrid Approach

Abstract

Foreground detection is one of the fundamental preprocessing steps in many image processing and computer vision applications. In spite of significant efforts, however, slowly moving foregrounds or temporarily stationary foregrounds remains challenging problem. To address these problems, this paper presents a hybrid approach, which combines background segmentation and long-term tracking with selective tracking and reducing search area, we robustly and effectively detect the foreground objects. The evaluation of realistic sequences from i-LIDS dataset shows that the proposed methodology outperforms with most of the state-of-the-art methods.