Internet access services over wireless networks have already been widely used by cellular phone users in Japan. However users have difficulty using keypads to browse web contents. To reduce this difficulty and offer smooth Internet access via mobile web browser, we have developed a voice input system that works on existing Internet services over cellular phones. This system works based on a combination of speech recognition by circuit switching and Internet access by packet switching. Through the field trial of this system, it was found that the speech input to the system contains various kinds of non-stationary noises. These type of noises often cause serious recognition errors especially in background speech noises. To reduce these errors, we propose a keyword spotting method using a garbage model for background speech to improve discrimination between speech and background speech noises that suffer from the characteristic distortion caused by low bit rate speech CODEC. Experiments in recognition of noisy speech show that our proposed method reduces word errors by 40% compared with results from speech recognition without a garbage model.
Cite as: Naito, M., Fujita, K., Shimizu, T. (2004) A noise robust voice input system for internet services over cellular phones. Proc. 9th Conference on Speech and Computer (SPECOM 2004), 231-235
@inproceedings{naito04_specom,
author={Masaki Naito and Kengo Fujita and Tohru Shimizu},
title={{A noise robust voice input system for internet services over cellular phones}},
year=2004,
booktitle={Proc. 9th Conference on Speech and Computer (SPECOM 2004)},
pages={231--235}
}