LibSonar: An Speech Indexing Based Chinese History Audio-Library

Along with the development of the Internet, people are increasingly dependent on acquiring knowledge and information from the Internet. This information includes text, audio, video and other abundant multimedia formats. Audio resource, as one friendly information carrier, is becoming more and more important and widespread on the Internet. It has become an important component of digital libraries. Audio-Library, which combines txt and audio resources, is an emerging network service. However, these services only retrieve audio resource basing on document names or meta information, not deeply on the content. In these organizations, resources are isolated and not correlative to read. The system is closed to users, on which general users can not easily submitt resources and the feedback.Comparing and investigating several related existing systems, we integrate the speech keyword recognition technology and digital library technology and propose an open framework of audio-library system based on audio-retrieval. This system can effectively retrieve text and audio information, support the associative search between information and is open to users, who could revise searches, add resources and increase indexes to the system. Addressing the issue of Chinese history and culture, LisSonar system is implemented in this thesis, which provides inquiry service of historical information and audio resource.This thesis mainly focuses on the following work:1. Proposing a new Audio-Library System framework and descripting its operation mechanism.2. Applying Speech Keywords Recognition technology in Audio-Library system.3. Applying inverted index in associated index and audio-keyword Index.4. Implementing a demo system focusing on Chinese historical information retrieval.The work is supported by the following foundings: NCET-04-0545, NSFC₆0525202/60533040, 863 Program₂006AA01Z136, PCSIRT0652, ZPNSF_Y106705.