Font Size: a A A

Research And Implementation Of HMM - Based Chinese Command Recognition System

Posted on:2017-05-07Degree:MasterType:Thesis
Country:ChinaCandidate:Z C TuFull Text:PDF
GTID:2278330485464388Subject:Engineering
Abstract/Summary:PDF Full Text Request
With the progress and development of science and technology, Smart phones, Smart home, as well as a series of smart products have entered our life,the concept of artificial intelligence has also become a hot research topic. Speech recognition is an important direction of artificial intelligence,which enables the computer to "understand" the language of the human beings,help people interact with the computer faster,better and more convenient.This paper makes a systematic study on speech recognition method based on Hidden Markov model(HMM),and we design and implement a speech recognition system with HMM theory. Then we use this system to complete a series of experiments, and verify the validity of the theory.The main contents are as followsing:(1)The paper introduces the basic structure of speech recognition system based on HMM,the preprocessing of the speech signal, MFCC feature parameter extraction, HMM training,the identification based on HMM and the basic principles of some other major modules.(2)On the basis of HMM speech recognition method, a command recognition system is designed and developed in this paper using C++ language,which introduces the main modules of the function and algorithm implementation.(3)A corpus is designed and collected according to the platform of speech recognition system developed in this paper.Then the entire system is tested to verify its effectiveness. Some experiments are carried out according to the different parameters and algorithms of the system, and the experimental data were collected and analyzed.(4) The Ada Boost algorithm is added to the system to improve the robustness of the system.
Keywords/Search Tags:Hidden Markov Model, Speech Recognition, Mel Frequency Cepstrum Coefficient
PDF Full Text Request
Related items