8th European Conference on Speech Communication and Technology

Geneva, Switzerland
September 1-4, 2003


Tree-Structured Noise-Adapted HMM Modeling for Piecewise Linear-Transformation-Based Adaptation

Zhipeng Zhang (1), Kiyotaka Otsuji (1), Sadaoki Furui (2)

(1) NTT DoCoMo Inc., Japan
(2) Tokyo Institute of Technology, Japan

This paper proposes the application of tree-structured clustering to various noise samples or noisy speech in the framework of piecewise-linear transformation (PLT)-based noise adaptation. According to the clustering results, a noisy speech HMM is made for each node of the tree structure. Based on the likelihood maximization criterion, the HMM that best matches the input speech is selected by tracing the tree from top to bottom, and the selected HMM is further adapted by linear transformation. The proposed method is evaluated by applying it to a Japanese dialogue recognition system. The results confirm that the proposed method is effective in recognizing noise-added speech under various noise conditions.

Full Paper

Bibliographic reference.  Zhang, Zhipeng / Otsuji, Kiyotaka / Furui, Sadaoki (2003): "Tree-structured noise-adapted HMM modeling for piecewise linear-transformation-based adaptation", In EUROSPEECH-2003, 669-672.