论文部分内容阅读
A formula to compute the similarity between two audio feature vectors is proposed, which can map arbitrary pair of vectors with equivalent dimension to [0,1). To fulfill the task of audio segmentation, a self-similarity matrix is computed to reveal the in