Feature-based detection and correction of occlusions and split of video objects

This paper proposes a novel algorithm for the real-time detection and correction of occlusion and split in object tracking for surveillance applications. The paper assumes a feature-based model for tracking and is based on the identification of sudden variations of spatio-temporal features of objects to detect occlusions and splits. The detection is followed by a validation stage that uses past tracking information to prevent false detection of occlusion or split. Special care is taken in case of heavy occlusion, when there is a large superposition of objects. For the detection of splits, in addition to the analysis of spatio-temporal changes in objects’ features, our algorithm analyzes the temporal behavior of split objects to discriminate between errors in segmentation and real separation of objects, such as in a deposit event. Both objective and subjective experimental results show the ability of the proposed algorithm to detect and correct, both, split and occlusion of objects. The proposed algorithm is suitable in video surveillance applications due to its good performance in multiple, heavy, and total occlusions, its ability to differentiate between real object separation and faulty object split, its handling of simultaneous occlusion and split events, and its low computational complexity. The algorithm was integrated into an on-line video surveillance system and tested under several conditions with promising results.