Takagi KAZUYUKI (高木 一幸)

Extraction of features from input speech that are effective in distinguishing the language is a key issue for language identification system. We use posterior probabilities on articulatory classes as features for language identification. Posterior probability on each articulatory class is calculated by GMMs. Each GMM is trained with MFCC data of speech segments labeled with the phonemes or acoustic events that correspond to the articulatory class. The posterior probability values of the articulatory classes are concatenated to form an articulatory-feature- class-posterior-probability (AFCPP) vector at each analysis frame. These vectors are then quantized to yield VQ code sequence, which is used as the training data for a n-gram language model. Language identification is performed by selecting the n-gram model that yields the highest likelihood for the AFCPP vector sequence of the input utterance. Language identification experiment between Japanese and English by the present method showed identification rate of 97.1%.

Language identification is the technique to identify the language being spoken by an unknown speaker. In this paper, phonotactic information was used as the feature for language identification. In order to obtain phonotactic information, it is required to extract the phoneme sequence from speech data. A template-based non-negative matrix factorization was applied for this purpose. The extracted phoneme sequence was then analyzed to yield n-gram models which may reflect the order in which the phoneme-like categories of speech occur in the language. Language identification was carried out by a support vector machine with the n-gram as the feature vector. It is shown that the identification performance changes with the number of spectrum templates and the order of n-gram, and that the best performance of 98.6% was obtained when the number of spectrum was 13 and the order of n-gram was 3.