Techniques for Noise Robustness in Automatic Speech Recognition

自动语音识别之噪声鲁棒性技巧

通信技术

售   价:
1091.00
发货周期:预计3-5周发货
作      者
出  版 社
出版时间
2012年10月30日
装      帧
精装
ISBN
9781119970880
复制
页      码
514
语      种
英文
综合评分
暂无评分
我 要 买
- +
库存 30 本
  • 图书详情
  • 目次
  • 买家须知
  • 书评(0)
  • 权威书评(0)
图书简介
Techniques for Noise Robustness in Automatic Speech Recognition will comprise chapters covering the state-of-the-art techniques written by experts in the field. Model compensation and adaptation will be considered which temporarily modify the statistical parameters employed by the recognizer to improve recognition of corrupted data. Signal compensation and signal separation methods attempt to reduce the level of noise in noisy speech. Feature compensation methods attempt to reduce the effect of corrupting noise on features derived from the speech. Missing feature methods concentrate on information-carrying components of the speech signal. Noise-robust feature estimation methods derive fundamentally noise-robust features from speech for recognition. Other methods deal optimally with packet loss, detection of bursty noises and other such phenomena. The book will comprise representative chapters that cover the state-of-the-art in all of the above approaches.
馆藏图书馆
Harvard Library
本书暂无推荐
本书暂无推荐
看了又看
  • 上一个
  • 下一个