Short-Term Motion-Based Object Segmentation

Supplementary material for:

"Short-Term Motion-Based Object Segmentation"

Marina Georgia ArvanitidouMichael TokAndreas Krutz and Thomas Sikora
IEEE International Conference on Multimedia & Expo (ICME), Barcelona, Spain, 11.07.2011 - 15.07.2011 


Motion-based segmentation approaches employ either long-term motion information or suffer from lack of accuracy and robustness. We present an automatic motion-based object segmentation algorithm for video sequences with moving camera, employing short-term motion information solely. For every frame, two error frames are generated using motion compensation. They are combined and a thresholding segmentation algorithm is applied. Recent advances in the field of global motion estimation enable outlier elimination in the background area, and thus a more precise definition of the foreground is achieved. We propose a simple and effective error frame generation and we consider spatial error localization. Thus, we achieve improved performance compared with a previously proposed short-term motion-based method and we provide subjective as well as objective evaluation.

Allstars (352 x 288, 250 frames)

Results and f-measure   
Allstars original---
Allstars reference180%48%59%
Allstars proposed73%66%68%

Biathlon (352 x 288, 200 frames)

Results and f-measure   
Biathlon original---
Biathlon reference182%69%74%
Biathlon proposed75%85%80%

Mountain (352 x 192, 100 frames)

Results and f-measure   
Mountain original---
Mountain reference192%73%81%
Mountain proposed76%91%83%

Race (544 x 336, 100 frames)

Results and f-measure   
Race original---
Race reference186%59%68%
Race proposed63%89%74%

Stefan (352 x 240, 300 frames)

Results and f-measure   
Stefan original---
Stefan reference190%53%63%
Stefan proposed69%82%73%


1 R. Mech, and M. Wollborn, “A noise robust method for segmentation of moving objects in video sequences”, in Proceedings of the IEEE International Conference on Acoustics, Speech and Signal Processing, 1997