Approximated Posterior Probability for Scoring Speech Recognition Confidence

다국어 초록 (Multilingual Abstract)

This paper proposes a new confidence measure for utterance verification with posterior probability approximation. The proposed method approximates probabilistic likelihoods by using Viterbi search characteristics and a clustered phoneme confusion matrix. Our measure consists of the weighted linear combination of acoustic and phonetic confidence scores. The proposed algorithm shows better performance even with the reduced computational complexity than those utilizing conventional confidence measures.

참고문헌 (Reference)

1 "Word and phone level acoustic confidence scoring" 1799-1820, 2000

2 "Vocabulary independent discriminative utterance verification for nonkeyword rejection in subword based speech recognition" 4 (4): Nov.1996

3 "Meta-models for confidence estimation in speech recognition" 1815-1818, 2000

4 "Implementation of the POW (phonetically optimized words) algorithm for speech database" 89-92, 1995.

5 T. Kemp, "Estimating confidence using word lattices" Greece 827-830, Sept.1997

6 "Confidence measures for large vocabulary continuous speech recognition" 9 (9): Mar.2001

7 "A probabilistic approach to confidence estimation and evaluation" 879-882, 1997

8 "A new approach to utterance verification based on neighborhood information in model space" 11 (11): Sept.2003