Automatic evaluation of children's performance on an English syllable blending task

In this paper, speech recognition techniques are applied to automatically evaluate children's performance in a syllable blending task. Word verification is performed to filter out utterances pronounced incorrectly. For valid words, forced alignment is applied to generate syllable segmentations and produce the corresponding HMM log likelihood scores. Normalized spectral likelihoods and duration ratio scores are combined to assess the overall quality of children's productions. Speaker-specific information is further incorporated to optimize performance. Experimental results show that the automatic system correlates well with those of teachers, but requires no human supervision.

[1]  Mitch Weintraub,et al.  Automatic text-independent pronunciation scoring of foreign language student speech , 1996, Proceeding of Fourth International Conference on Spoken Language Processing. ICSLP '96.

[2]  Roger Sensenbaugh,et al.  Phonemic Awareness: An Important Early Step in Learning To Read. ERIC Digest. , 1996 .

[3]  Yoon Kim,et al.  Automatic pronunciation scoring for language instruction , 1997, 1997 IEEE International Conference on Acoustics, Speech, and Signal Processing.

[4]  Rodolfo Delmonte,et al.  SLIM prosodic automatic tools for self-learning instruction , 2000, Speech Commun..

[5]  Fabio Tamburini,et al.  Prosodic prominence detection in speech , 2003, Seventh International Symposium on Signal Processing and Its Applications, 2003. Proceedings..

[6]  Abeer Alwan,et al.  TBALL data collection: the making of a young children's speech corpus , 2005, INTERSPEECH.