David Grangier, Alessandro Vinciarelli
This paper presents experiments that evaluate the effect of different video segmentation methods on text-based video retrieval. Segmentations relying on modalities like speech, video and text or their combination are compared with a baseline sliding window ...
IDIAP2004