Adaptive Video Partitioning
(M. C. Lee)

Many different algorithms have been proposed for video partitioning; however, their performance in terms of precision and recall varies with the characteristics of the video sequence. There is no single algorithm, which can detect scene breaks satisfactorily for different video sequences. This project attempts to classify different types of video sequences and to develop an adaptive approach to the partitioning of video scenes with the aim to improve overall performance.