Research results on background subtraction

We present a new algorithm for background modeling and subtraction. Sample background values at each pixel are quantized into codebooks which represent a compressed form of background model for a long image sequence. This allows us to capture structural background variation due to periodic-like motion over a long period of time under limited memory. Our method can handle scenes containing moving backgrounds or illumination changes, and it achieves robust detection for compressed videos.

1. Detection on compressed video

2. Layered detection

3. Moving trees

4. Unattended objects