Font Size: a A A

Research And Development Of Crying Detection System Based On Machine Learning

Posted on:2019-12-04Degree:MasterType:Thesis
Country:ChinaCandidate:X W MaoFull Text:PDF
GTID:2428330566496123Subject:Electronic and communication engineering
Abstract/Summary:PDF Full Text Request
In the era of smart home,smart products are highly favored by users,reflecting the urgent need for innovative services.It is an effective way to realize the innovation and performance improvement on traditional hardware devices and services by use of machine learning technology.Embed the function of baby care in smart products,can detect baby crying in time,and alert the user to avoid danger when the baby is alone,it has greater market application value.Based on this,this paper uses cry detection technology to monitor baby cries,and then researches and develops intelligent cry detection system based on machine learning,the work of this article is briefly described as follows.?.Research on typical machine learning algorithms in voiceprint recognition system: study the algorithm principles of each machine learning model,and simulate the GMM and SVM models to realize the cry signals by comparing the two indicators of accuracy and speed.Recognition performance,and provide a theoretical basis for the following research based on the hybrid model-based intelligent cry detection technology.?.Research on key technical solutions for intelligent cry detection system: First,a demand analysis is performed for the intelligent cry detection system,Secondly,the hardware and software architecture diagram of the intelligent cry recognition system is designed.finally,two key technologies in the cry detection system were proposed: “multi-feature extraction technology based on Mel energy” and “intelligent cry detection technology based on hybrid model” as intelligent cries.The research and development of the detection system provides an solution.?.Developed and implemented a machine learning-based intelligent cry detection system.Firstly,it elaborates the specific implementation details of the mobile terminal software,intelligent hardware and software of the embedded device,and the model training process in the intelligent cry detection system.Secondly,the system is tested in function and performance.Strict system testing verifies the advantages of technical and performance.
Keywords/Search Tags:Machine learning, Cries detection, Voiceprint feature, Voiceprint modeling, System development
PDF Full Text Request
Related items