Study On Content-based Audio Retrieval Technology

Multimedia information retrieval has become one of the most active research areas in attracted people in the past decades. And in tile beginning most work were concentrated on content-based image retrieval (CBIR) and content-based video retrieval(CBVR). However, as there are more and more digital audio data in place at present, people start to realize the importance of content — based audio retrieval. With tile development of some related key technologies including audio signal process, speech recognition, speaker recognition, and music automatic transcription, etc, it has established the basis of content-based audio retrieval (CBAR).Content-based audio retrieval has a wide range of applications in distance education medical treatment, digital library, environment detection, news program retrieval, entertainment industry, and surveillance, and so on. Many researches have been done on content-based audio classification and segmentation, spoken document retrieval (SDR), music retrieval, audio content description and indexing, and query construction and refinement. Incorporating these research fruits, this dissertation deals with automatic audio content analysis that solves the following problems, such as content -based audio classification.The major content of this dissertation can be concluded as follows: (1) Audio Data Pattern;(2) The Basis Theoretical of Hidden Markov Model ( HMM) ;(3) Research on Hidden Markov Model based Automatic Audio Classification Algorithm.