Background Modeling Techniques in Scene Surveillance System

Pixelwise Median filter:
Median filtering is one of the most commonly-used background modeling techniques [1,2,3].The background estimate is denoted to be the median at each pixel location of all the frames in the buffer.The assumption is that the pixel stays in the background for more than half of the frames in the buffer.Therefore, it is applied to several seconds of video to distinguish moving pixels from stationary pixels.
Median filtering has been extended to color by replacing the median with the medoid algorithm [2]. The complexity of computing the median is O(LlogL) for each pixel.

Reference:

• [1] Cutler, R.; Davis, L. "View-based detection and analysis of periodic motion", in Proceedings Fourteenth International Conference on Pattern Recognition, 1, pp. 495-500, (Brisbane, Australia), Aug 1998.
• [2] Haritaoglu, I.; Harwood, D.; Davis, L.S. " W4: real-time surveillance of people and their activities", Pattern Analysis and Machine Intelligence, IEEE Transactions on Volume 22, Issue 8, pp. 809 - 830, Aug. 2000
• [2] Cucchiara, R.; Grana, C.; Piccardi, M.; Prati, A. "Detecting moving objects, ghosts, and shadows in video streams", Pattern Analysis and Machine Intelligence, IEEE Transactions on Volume 25, Issue 10, pp. 1337 - 1342, Oct. 2003