摘要
为了提高不同噪音环境中智慧家居的声纹识别效果,本研究以深度学习为基础,旨在优化并提高智慧家居混合声纹识别率。采用改进的混合声纹识别系统,在不同噪声环境下进行实验。结果显示,相较于传统方法,本研究改进的混合声纹识别系统在整体性能上提高了15%以上,能够精确识别智慧家居控制者的声纹信息与特征。特别是在厨卫噪音中,识别精度达到了90%以上。因此,本研究所构建的混合声纹识别系统在厨卫相关的声纹识别领域具有潜在应用价值,并在其他噪音环境下同样表现良好。
In order to improve the voiceprint recognition performance of smart homes in different noise environments,this study is based on deep learning and aims to optimize and improve the mixed voiceprint recognition rate of smart homes.Using an improved hybrid voiceprint recognition system,experiments were conducted in different noise environments.The results show that compared to traditional methods,the improved hybrid voiceprint recognition system in this study has improved overall performance by more than 15%,and can accurately recognize the voiceprint information and features of smart home controllers.Especially in kitchen and bathroom noise,the recognition accuracy has reached over 90%.Therefore,the hybrid voiceprint recognition system constructed in this study has potential application value in the field of kitchen and bathroom related voiceprint recognition,and performs well in other noisy environments.
作者
张晶晶
李晨鸣
ZHANG Jingjing;LI Chenming(Xianyang vocational and technical college,Xianyang Shanxi 712000,China;Xianyang First People’s Hospital,Xianyang Shanxi 712000,China)
出处
《自动化与仪器仪表》
2023年第12期139-143,共5页
Automation & Instrumentation
基金
咸阳职业技术学院专题研究项目《“双高”背景下创新创业教育课程体系的构建与研究》(2022ZYA03)。
关键词
智慧家居
噪音
高斯混合模型
循环神经网络
高精度识别
smart home
noise
gaussian mixture model
recurrent neural network
high precision recognition