Abstract
This paper proposes a method to automatically determine the sports type of a sports game based on KWS (keyword spotting) techniques. First, we develop an audio segmentation module as the front-end to extract announcer’s speech efficiently from the complex sports audio stream. Then we employ speech recognition technology on these speech segments to extract keywords as the features of each kind of sports. Finally, based on the improved KWS results and specific keywords selected for each kind of sports, the classification is conducted based on a vote ranking strategy. For robust KWS in our system, adaptation techniques for acoustic model and language model are employed. In the acoustic model adaptation, supervised adaptation is carried out using MAP(maximum a posterior). In the language model adaptation, a keyword-frequency-based adaptation is proposed in this paper. Both adaptations show significant improvements on KWS performance. By integrating all the techniques, we achieve 100% accuracy rate in STD (sports type determination) tested on 15 games of seven kinds of sports.
Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
