Two-stage content-based audio segmentation algorithm

Content-based audio segmentation plays an important role in multimedia applications. In order to segment accurately and on-line, most conventional algorithms are based on small-scale audio classification and always result in a high false segmentation rate. The authors' experimental results show that large-scale audio can be more easily classified than small ones, and this trend is irrespective of classifiers. According to this fact, this paper presents a novel framework for audio segmentation to reduce the false segmentations. First, a rough segmentation step based on large-scale audio classification is taken to ensure the integrality of the content of audio segments, which can avoid the consecutive audio belonging to the same kind being segmented into different pieces. Then a subtle segmentation step based on segmentation point evaluation function is taken to further locate the segmentation points for the boundary areas computed by the rough segmentation step. Experimental results show that nearly 3/4 false segmentation points can be reduced comparing to the conventional audio segmentation method based on small-scale audio classification, while preserving a low missing rate.