Mining TV broadcasts for recurring video sequences
Proceedings of the ACM International Conference on Image and Video Retrieval
TV program segmentation using multi-modal information fusion
Proceedings of the 1st ACM International Conference on Multimedia Retrieval
| Hi-index | 0.00 |
Three different frame features (color patches, color co- herence vectors, and gradient histograms) are investigated for their suitability to recognize recurring video clips in very large databases. They are evaluated in a real-time pro- cessing and real-time recognition system. Real-time recog- nition means that each clip must be recognized one second after its start. As the experimental results show, only gra- dient histograms work satisfactorily across different video material with the same video domain independent param- eter set. For instance, they are in contrast to color fea- tures not negatively affected by dark frame sequences in video clips and the live video stream. By means of pre- computation and subsequent table look-ups, gradient his- tograms can be implemented such that their computational costs come very close to that of color features.