Real-Time Voice Command Recognition System Based on DTW and MFCC

Voice command recognition is an important research direction in the field of intelligent speech. Its task is to detect and recognize a given set of command words in continuous speech streams. For command recognition tasks with a voice command vocabulary, this article establishes a real-time voice command recognition system based on DTW and MFCC, using Python tools, to meet the requirements of real-time command recognition tasks. Firstly, the system processes the real-time collected voice based on dynamic time warping (DTW), and then uses the Deep Neural Network (DNN) as a feature extractor to extract the MFCC feature vector of the voice, and calculate the Euclidean distance between the feature vectors of the test voice and the template. The experimental results demonstrate that the real-time speech command recognition system based on DTW and MFCC has comprehensive performance advantages, which can meet the application scenarios of real-time control of speech commands and free users' hands.