Research Studies Project Report (M.Eng.) - Asian Institute of Technology, 2002

A research study submitted in partial fulfillment of the requirement for the degree of Master of Engineering, School of Advanced Technologies

The increased availability and usage of multimedia information stored in digital form have created a critical need for efficient multimedia processing algorithms. These algorithms must offer capabilities related to browsing, indexing and retrieval of relevant data. The process of automatic shot boundary detection is a fundamental component in video indexing, editing and archiving. The performance of video indexing and retrieval system depends on the results of scene change detections. This research-work studies video compression technique, MPEG-4 coding standard and gradual scene transition detection in the MPEG-4 compressed video domain. The MPEG-4 object-based coding standard, designed as a common platform for all multimedia applications, is inherently well suited for video indexing applications. In this research-work we implement an algorithm for real-time detection of fadein/ fade-out scene transitions using texture and macroblock information in the MPEG-4 compressed video. The performance of the proposed algorithm is measured in terms of precision and recall values. The experiments show that hard cut and fading transitions can be detected reliably, if the threshold value of shot detection is selected appropriately to the input video type.