Improving the Performance of Acoustic Event Classification by Selecting and Combining Information Sources Using the Fuzzy Integral
- 1.3k Downloads
Acoustic events produced in meeting-room-like environments may carry information useful for perceptually aware interfaces. In this paper, we focus on the problem of combining different information sources at different structural levels for classifying human vocal-tract non-speech sounds. The Fuzzy Integral (FI) approach is used to fuse outputs of several classification systems, and feature selection and ranking are carried out based on the knowledge extracted from the Fuzzy Measure (FM). In the experiments with a limited set of training data, the FI-based decision-level fusion showed a classification performance which is much higher than the one from the best single classifier and can surpass the performance resulting from the integration at the feature-level by Support Vector Machines. Although only fusion of audio information sources is considered in this work, the conclusions may be extensible to the multi-modal case.
KeywordsSupport Vector Machine Hide Markov Model Information Source Support Vector Machine Classifier Automatic Speech Recognition
Unable to display preview. Download preview PDF.
- 1.Temko, A., Nadeu, C.: Meeting room acoustic event classification by support vector machines and variable-feature-set clustering. In: ICASSP 2005, Philadelphia (March 2005)Google Scholar
- 3.Weston, J., Mukherjee, J., Chapelle, O., Pontil, M., Poggio, T., Vapnik, V.: Feature Selection for SVMs. In: Proc. of NIPS (2000)Google Scholar
- 4.Sugeno, M.: Theory of fuzzy integrals and its applications, PhD thesis, Tokyo Institute of Technology (1974)Google Scholar
- 5.Grabisch, M.: The Choquet integral as a linear interpolator. In: 10th Int. Conf. on Information Processing and Management of Uncertainty in Knowledge-Based Systems (IPMU 2004), Perugia (Italy), July 2004, pp. 373–378 (2004)Google Scholar
- 8.Chang, S., Greenberg, S.: Syllable-proximity evaluation in automatic speech recognition using fuzzy measures and a fuzzy integral. In: Proc. of the 12th IEEE Fuzzy Systems Conf., pp. 828–833 (2003)Google Scholar
- 9.Grabisch, M.: A new algorithm for identifying fuzzy measures and its application to pattern recognition. In: Proc. of 4th IEEE Int. Conf. on Fuzzy Systems, Yokohama, Japan, pp. 145–150 (1995)Google Scholar
- 10.Wu, Y., Chang, E., Chang, K., Smith, J.: Optimal Multimodal Fusion for Multimedia Data Analysis. In: Proc. ACM Int. Conf. on Multimedia, New York, pp. 572–579 (October 2004)Google Scholar
- 11.Kuncheva, L.: Combining classifiers: Soft computing solutions. Lecture Notes in Pattern Recognition, pp. 427–452. World Scientific Publishing Co, Singapore (2001)Google Scholar
- 14.Kojadinovic, I., Marichal, J.-L., Roubens, M.: An axiomatic approach to the definition of the entropy of a discrete choquet capacity. In: 9th Int. Conf. on Information Processing and Management of Uncertainty in Knowledge-Based Systems (IPMU 2002), Annecy (France), pp. 763–768 (2002)Google Scholar
- 15.Evaluation Packages for the First CHIL Evaluation Campaign. In: CHIL project Deliverable D7.4 (Mar 2005) Available at, http://chil.server.de/servlet/is/2712/
- 18.Nadeu, C., Hernando, J., Gorricho, M.: On the decorrelation of filter-bank energies in speech recognition. In: Proc. Eurospeech 1995, pp. 1381–1384 (1995)Google Scholar