说话人转变检测的分析与实现-analysis and implementation of speaker transition detection.docxVIP

Ⅱ Ⅱ Abstract With the development of IT technology, rich access to varies type of audio document and the growth of data volume, it is more and more difficult to manage the document. In recent years, audio segmentation and clustering technology have been researched to deal multimedia speech document, the most difficult one is meeting voice. Segmentation and clustering based on speakers is to distinguish voices of speakers and segment speeches into many segmentations in which everyone contains only one speaker, then mark the same speaker and reset with speaker clustering after segmentation. Segmentation is completed primarily with Speaker change detection (SCD), which is to find change time between two different speakers. The segmentation and clustering proposed in the paper consist of three aspects: feature extraction, speaker segmentation and the speaker similarity detection. They are introduced in detail and advantages and disadvantages of different methods are listed with experiments. contents are as follows: Speech feature extraction. LPCC and MFCC are applied for speaker characteristic parameters, through the experiment finds MFCC performance better than LPCC. Speaker segmentation. The detection method based on mixed speaker change with credibility trend and improved BIC is to utilize credibility trend to solve cumulative error caused by data accumulated and to utilize BIC to solve error caused by improper credibility parameters percentage. The experimental results show that the hybrid algorithm increase by 10% and 5.8% respectively than their individual use. Speaker similarity detection. The paper proposed that gender recognition based on pitch period and formant, speaker similarity detection and clustering based on GMM model. The proposal is verified to apply to occasions with a small number of people, such as telephone dialog and small meeting. Keyword: speaker change detection; feature extraction; speaker segmentation; speaker clustering Ⅳ Ⅳ 目 录 HYPERLINK \