Short description of Asahi-kasei's SBD system on TRECVID 2007

1. Briefly, what approach or combination of approaches did you test in each of your submitted runs? (please use the run id from the overall results table NIST returns) * (runid: a sentence or series of keywords characterizing run) * (runid: a sentence or series of keywords characterizing run) * ... sei1: Base system with cut and fade detection 2D correlation of successive frames in low resolution gray scale image is used to detect cuts. Together with Euclid's norm size to check reliability of correlation and detection of fades. sei2: additional gradual detection with additional long term 2D correlation in time(-50 and +50 fames) to detect gradual changes. 2. What if any significant differences (in terms of what measures) did you find among the runs? Lot of missing graduals for both runs. Additional gradual detectors helped for sum degree but was not enough. As we planed to add some detectors, did not adjust parameters for best results with lot of false graduals.