Robust and Accurate Segmentation of Moving Objects in Real-time Video

Robust and accurate segmentation of moving object in real-time video is very important for object silhouette extraction in vision-based human computer interaction and video surveillance systems. However, the inherent problem of moving object segmentation based on background subtraction is to distinguish the changes from background disturbing effects such as noise and illumination changes. Therefore, the paper proposes an improved background subtraction scheme which is robust and accurate against noisy and changing illumination. The occlusion regions are detected based on the frame difference and background difference images. The moving shadows are eliminated very effectively by using the background statistical parameters. A queue-based connected component analysis method is introduced to isolate the moving object from background noise. Moreover, a pixel based background update is used to update the illumination changes. The proposed scheme has been implemented and evaluated regarding the segmentation quality and frame rate. From the experimental results it is known that the proposed method successfully extract object contours accurately against the illumination and noise changes.