A holistic, in-compression approach to mining independent motion segments for massive surveillance video collections

Abstract

This chapter describes a large scale surveillance video data mining approach for those segments that contain independently moving targets. Given the typical scenario where the video data collections are massive in size, We propose a holistic, in-compression approach, called Linear System Consistency Analysis (LSCA), to efficient video data mining for those independent motion segments. By efficient, we mean that the mining speed is close to or even faster than real-time in "normal" platforms (we do not assume using special hardware or any parallel machines) while still maintaining a good mining quality. Theoretical and experimental analyses demonstrate and validate this holistic, in-compression approach to solving for video mining problem for temporal independent motion segmentation. © 2010 Springer-Verlag Berlin Heidelberg.