说话人识别完整源码（matlab实现）

# Project Description: - In this project, three major tasks are completed: - We implement the EM algorithm for a Gaussian Mixture Model for the task of speaker identification given a set of speech acoustics vectors (MFCC form). - We used the Viterbi algorithm on a continuous Hidden Markov Model to learn the latent phoneme sequences in a given a speech waveform - thereby allowing us to do speech recognition. We implement the Levenshtein distance algorithm to evaluate our speech recognition system. - We use the IBM BlueMix service in order to do speech synthesis - Developed by: - [Munir Al - Dajani](https://www.linkedin.com/in/munirad) - [Benjamin Lavon](https://www.linkedin.com/in/benjamin-lavon-1a09b6124) - Refer to [handout.pdf](https://github.com/MunirAD/Speaker_Identification_Speech_Recognition/blob/master/handout.pdf) for a more detailed explanation of this projects tasks. ## Description of Files - Refer to [handout.pdf](https://github.com/MunirAD/Speaker_Identification_Speech_Recognition/blob/master/handout.pdf) for a detailed description of the files