论文信息 - TREC-10 Shot Boundary Detection Task: CLIPS System Description and Evaluation

TREC-10 Shot Boundary Detection Task: CLIPS System Description and Evaluation

This paper presents the system used by CLIPS-IMAG to perform the Shot Boundary Detection (SBD) task of the Video track of the TREC-10 conference. Cut detection is performed by computing image difference after motion compensation. Dissolve detection is performed by the comparison of the norm over the whole image of the first and second temporal derivatives. An output filter is added in order to clean the data of both detector and to merge them into a consistent output. This system also has a special module for detecting photographic flashes and filtering them as erroneous cuts. Results obtained for the TREC-10 evaluation are presented. The system appear to perform in a very good way for cut transitions and within the average for gradual transitions.

Georges Quénot

[1] Mohan S. Kankanhalli,et al. Temporal multiresolution analysis for video segmentation , 1999, Electronic Imaging.

[2] Philippe Aigrain,et al. Medium knowledge-based macro-segmentation of video into sequences , 1997 .

[3] Rosa Ruiloba. Towards a standard protocol for the evaluation of temporal video segmentation algorithms , 1999 .

[4] QU GeorgesM. Computation of Optical Flow using Dynamic Programming , 1996 .

[5] John S. Boreczky,et al. Comparison of video shot boundary detection techniques , 1996, Electronic Imaging.

[6] Dragutin Petkovic,et al. Content-based representation and retrieval of visual media: A state-of-the-art review , 1996, Multimedia Tools and Applications.