FXPAL at TRECVID 2005

The shot boundary detection system we are using for 2005 builds on the framework and system developed in 2004 which combines pairwise similarity analysis and supervised classification. Using primitive lowlevel image features, we build secondary features based on inter-frame dissimilarity. These secondary features are used as input to an efficient k-Nearest-Neighbor (kNN) classifier. The classifier labels each frame as a shot boundary or non-boundary, and the classifier outputs are minimally processed to determine the final segmentation. This year we added information-theoretic feature selection to determine two secondary feature subsets to improve cut transition detection and gradual transition detection, respectively. These systems appeared as runs sys10M 0X in the run table. This indeed improved performance over our baseline runs (sys05 0X), and improved on the performance of a similar system using random projection for dimension reduction (sys10R 0X). Our performance was worse than anticipated, as our training data was not an accurate reflection of the test data in the case of video from LBC and CCTV. On the remaining videos, our performance was very good, and consistent with our training experiments. The realizability of this approach remains an open question.