Identification of Malay Stop Consonants Based on MFCC &Rasta PLP Features Using K-NN Classifier for Cued Speech Application

摘要: Phonological studies suggest that phoneme awareness in an early age through cued speech is reliable to measure the literacy skills and provide a strong language foundation for deaf children. This paper proposed phonemic-based recognition of Malay Phonemes according stop voicing. Eight consonants /p b t d t∫ dʒ k g/ preceding /a/ vowel are selected encode each combination as hand shape at specified position. Features extracted by using Mel -frequency Cepstral Coefficients (MFCC), MFCC with delta coefficients, coefficients Rasta PLP. Mean featured group samples were taken reduce frame dimensions features. These dimensionalities reduced features fed into k-Nearest Neighbors (k-NN) classifier classification. K-fold cross validation method used test reliability results. Experimental results show best identification rate 92.5% upon feature fusion sets MFCC, PLP., voltage divider circuit waveforms inverter output shown MATLAB Simulink software.