An error-tolerant algorithm in decoding module of Mandarin continuous speech recognition is examined to correct substitution, insertion and deletion errors in acoustic recognition.
An error-tolerant algorithm in decoding module of Mandarin continuous speech recognition is examined to correct substitution, insertion and deletion errors in acoustic recognition.
摘要本文研究了汉语连续语音识别音字转换中的容错算法，以纠正声学识别的替代、插入、删除错误。