Research And System Implementation On Content Based Music Information Retrieval

Music is an important kind of audio data. The traditional music information retrieval system is almost based on metadata, and the use is limited. With the rapid development of the Internet and the popularization of digital devices, rapid and effective searching for desired audio data becomes an important and challenging research topic.In this thesis, considering the background and the recent studies on music retrieval, we study the main issues of content-based music information retrieval (MIR), on this basis we design and implement a query-by-semantic-text MIR prototype system.The main issues we study include the representation and extractation of music content, music signal processing and the methods and technologies on content-based MIR. By using SML model for reference which is proved to have achieved good effectiveness in the recent research on content-based image and music information retrieval, we present Chinese music retrieval based SML model, expressly create a Chinese music dataset PDM500 and represent the related features. On this basis, we design and implement a query-by-semantic-text MIR prototype system. The system is able to realize the function of music retrieval and annotation by learning the model .The evaluation shows an encouraging result ,however, some problems are found to be solved in the future work.The study work of this thesis provides the preparation for the developing of content based MIR research and pushes forward the study on semantic text retrieval.