TOWARDS AN OPTIMAL FEATURE SET FOR ENVIRONMENTAL SOUND RECOGNITION

Feature selection for audio retrieval is a non-trivial task . In this paper we aim at identifying an optimal feature combination for environmental sound recognition. The feature combination is constructed from a broad set of features. Additionally to state-of-the-art features, we evaluate the qua lity of audio features we previously introduced for another domain. We examine the properties of features by quantitative data analysis (factor analysis) and identify candidates fo r feature combinations. We verify the quality of the combination by retrieval experiments. The optimal solution yields Reca ll and Precision values of 87% and 88%, respectively.

[1]  Kiyoharu Aizawa.  Digitizing Personal Experiences: Capture and Retrieval of Life Log , 2005, MMM.

[2]  Christian Breiteneder,et al.  Discrimination and retrieval of animal sounds , 2006, 2006 12th International Multi-Media Modelling Conference.

[3]  Yi-Ping Phoebe Chen,et al.  The power of play-break for automatic detection and browsing of self-consumable sport video highlights , 2004, MIR '04.

[4]  Tsuhan Chen,et al.  Audio feature extraction and analysis for scene classification , 1997, Proceedings of First Signal Processing Society Workshop on Multimedia Signal Processing.

[5]  Seung-Yong Chun,et al.  Acoustic intruder detection system for home security , 2005, IEEE Transactions on Consumer Electronics.

[6]  Surya Nepal,et al.  Automatic detection of 'Goal' segments in basketball videos , 2001, MULTIMEDIA '01.

[7]  Guodong Guo,et al.  Content-based audio classification and retrieval by support vector machines , 2003, IEEE Trans. Neural Networks.