For Full-Text PDF, please login, if you are a member of IEICE,|
or go to Pay Per View on menu list, if you are a nonmember of IEICE.
Noise and Channel Distortion Robust ASR System for DARPA SPINE2 Task
Konstantin MARKOV Tomoko MATSUI Rainer GRUHN Jinsong ZHANG Satoshi NAKAMURA
IEICE TRANSACTIONS on Information and Systems
Publication Date: 2003/03/01
Print ISSN: 0916-8532
Type of Manuscript: Special Section PAPER (Special Issue on Speech Information Processing)
Category: Robust Speech Recognition and Enhancement
noise robustness, online adaptation, hypothesis combination, robust features,
Full Text: PDF(457.6KB)>>
This paper presents the ATR speech recognition system designed for the DARPA SPINE2 evaluation task. The system is capable of dealing with speech from highly variable, real-world noisy conditions and communication channels. A number of robust techniques are implemented, such as differential spectrum mel-scale cepstrum features, on-line MLLR adaptation, and word-level hypothesis combination, which led to a significant reduction in the word error rate.