作者: Fang Zheng , Zhanjiang Song , Wenhu Wu , Mingxing Xu
DOI:
关键词:
摘要: The Speaking Skill Evaluation (SSE) technologies are derived from speech recognition and used for language learning instructing. In this paper, an effective automatic pronunciation scoring method SSE systems is proposed. Center-Distance Continuous Probability Model (CDCPM) incorporated to model the speech. Merging-Based Syllable Detection Automaton (MBSDA) Non-Linear Partition (NLP) perform time alignment. And Critical Area Percentage (CAP) based score learner’s pronunciations or reject invalid utterances. Subjective assessments show that concise, fast, effective. system on it has achieved a satisfying performance.