Deller J R, Hsu D, Ferrier L J
Michigan State University, Department of Electrical Engineering, East Lansing 48824.
Comput Methods Programs Biomed. 1991 Jun;35(2):125-39. doi: 10.1016/0169-2607(91)90071-z.
Recognition of the speech of severely dysarthric individuals requires a technique which is robust to extraordinary conditions of high variability and very little training data. A hidden Markov model approach to isolated word recognition is used in an attempt to automatically model the enormous variability of the speech, while signal preprocessing measures and model modifications are employed to make better use of the existing data. Two findings are contrary to general experience with normal speech recognition. The first is that an ergodic model is found to outperform a standard left-to-right (Bakis) model structure. The second is that automated clipping of transitional acoustics in the speech is found to significantly enhance recognition. Experimental results using utterances of cerebral palsied persons with an array of articulatory abilities are presented.
识别严重构音障碍者的语音需要一种对高度可变且训练数据极少的特殊情况具有鲁棒性的技术。本文采用隐马尔可夫模型方法进行孤立词识别，试图自动对语音的巨大变异性进行建模，同时采用信号预处理措施和模型修改以更好地利用现有数据。有两个发现与正常语音识别的一般经验相反。第一个发现是，遍历模型的性能优于标准的从左到右（巴克斯）模型结构。第二个发现是，语音中过渡声学特征的自动裁剪能显著提高识别率。本文展示了使用具有一系列发音能力的脑瘫患者话语的实验结果。
