The implementation of object-based shot boundary detection using edge tracing and tracking

With the upcoming development of MPEG-7, research in shot boundary detection has become more popular. Traditionally, shot boundary detection was done by measuring the differences between the features of adjacent frames and threshold was constructed based on a trial and error basis. Such a method does not work well under gradual transition. In this paper, we describe a new object-based detection, which utilizes information from multiple frames through a time stamp transferring mechanism. The concept of edge object tracker is developed to track fast moving objects as well as allowing the unbounded object to be traced across the gradual transition scenes. Our result shown that object-based detection is more superior to traditional indicators for detecting gradual transitions. Such technique works very well with frames containing skipped blocks and is independent of the transition period, which cannot be handled by indicators based on two-frame comparison.