Pattern Recognition, International Conference on (2000)
Sept. 3, 2000 to Sept. 8, 2000
Ig-Tae Um , Sungkyunkwan University
Jong-Hei Ra , Sungkyunkwan University
Moon-Hyun Kim , Sungkyunkwan University
This paper compares two clustering methods: SOM and a graph-based clustering technique for MLP-based text independent speaker verification. The focus of comparison is given to the distribution characteristics of representative frames for each cluster and to the processing time of clustering and MLP learning, and to verification performance. Simulation results show that the graph-based technique produces better verification performance than SOM. Other statistics are collected to explain significant difference in MLP learning time with each clustering method. This experiment suggests that there is a best match between a classifier and a clustering method for a given application.
J. Ra, M. Kim and I. Um, "Comparison of Clustering Methods for MLP-Based Speaker Verification," Pattern Recognition, International Conference on(ICPR), Barcelona, Spain, 2000, pp. 2475.