No abstract

No abstract

Abstract. This paper presents a novel surveillance video indexing and retrieval system based on object features similarity measurement. The system firstly extracts moving objects from the videos by an efficient motion segmentation method. The fundamental features of each moving object are then extracted and indexed into the database. During retrieval, the system matches the query with the features indexed in the database without re-processing the videos. Video clips which contain the objects with sufficiently high relevance scores are then returned. The novelty of the system includes: 1. A real-time automatic indexing methodology achieved by a fast motion segmentation, such that the system is able to perform on-the-fly indexing on video sources; and 2. an object-based retrieval system with fundamental features matching approach, which allows user to specify the query by providing an example image or even a sketch of the desired objects. Such an approach can search the desired video clips in a more convenient and unambiguous way comparing with traditional text-based matching.

This paper presents a robust human tracking system which incorporates automatic detection of head shape objects with decentralized tracking approach. A fast and robust probabilistic shape contour matching algorithm is applied to the input image frame to detect and locate head shape objects. The detected objects are then tracked by decentralized trackers. Here, a decentralized tracker refers to the tracker that tracks exactly one object. Essentially, each newly detected object will instantiate an individual tracker, which tracks the object and destroys itself when the object disappears. Two trackers communicate with each other only when they are getting close enough. This approach simplifies the competition of targets between trackers, and is more efficient than the centralized approach whose time complexity is greatly depends on the number of tracked objects. The system has been tested with several challenging digital surveillance video sequences, and the results show the robustness and the efficiency of the system under crowded and clutter environment.