A new approach for extracting and summarizing abnormal activities in surveillance videos

In this paper, we propose a new approach to detect abnormal activities in surveillance videos and create suitable summary videos accordingly. The proposed approach first introduces a blob sequence optimization process which integrates spatial, temporal, size, and motion correlation among objects to extract suitable abnormal blob sequences. With this process, blob extraction errors due to occlusion or background interferences can be effectively avoided. Then, we also propose an abnormality-type-based method which creates short-period summary videos for long-period input surveillance videos by properly arranging abnormal blob sequences according to their activity types. Experimental results show that our proposed approach can effectively create satisfying summary videos from input surveillance videos.