HMM-Based Digital Speech Recognition System

This HMM-based digital speech recognition system employs advanced algorithms and techniques to achieve accurate recognition of digits 0 through 9. The system implements a user-friendly GUI interface that enables convenient recording and playback operations through MATLAB's audio processing functions. Beyond the core digit recognition capability, the system includes preprocessing modules featuring endpoint detection and MFCC (Mel-Frequency Cepstral Coefficients) feature extraction to improve recognition accuracy. The HMM implementation utilizes Baum-Welch algorithm for training and Viterbi algorithm for decoding digit patterns. Additionally, the system comes with sample audio files allowing users to better understand the recognition results through practical demonstrations. Through this system, users can easily perform digital speech recognition tasks and obtain reliable results with robust code implementation handling feature extraction and pattern matching workflows.