中文核心期刊
中国科技论文统计源期刊
国际刊号:1004-9037
国内刊号:32-1367/TN
用户登录
  E-mail:  
  密  码:  
  作者 审稿  
  编辑 读者  
期刊向导
联系方式
  • 主管:中国科学技术协会
  • 主办:南京航空航天大学
  •           中国电子学会
  • 国际刊号:1004-9037
  • 国内刊号:32-1367/TN
  • 地址:南京市御道街29号
  • 电话:025-84892742
  • 传真:025-84892742
  • E-mail:sjcj@nuaa.edu.cn
  • 邮编:210016
基于图卷积深浅特征融合的跨语料库情感识别
Deep and shallow features fusion based on graph convolution for cross-corpus emotion recognition
投稿时间:2022-01-07  修订日期:2022-09-01
DOI:
中文关键词:  图卷积神经网络  跨语料库  语音情感识别  构图  深层和浅层特征融合
英文关键词:graph convolutional network  cross-corpus  speech emotion recognition  composition  deep and shallow features fusion
基金项目:江苏省高校自然科学基金项目
作者单位邮编
杨子秀 江苏师范大学 221116
金赟 江苏师范大学 221116
马勇 江苏师范大学 
戴妍妍 江苏师范大学 
俞佳佳 江苏师范大学 
顾煜 江苏师范大学 
摘要点击次数: 148
全文下载次数: 0
中文摘要:
      现实中,语音情感识别任务的训练数据和测试数据往往来源于不同的数据库,二者特征空间存在明显差异,导致识别率很低。针对该问题,本文提出新的构图方法表示源和目标数据库之间的拓扑结构,利用图卷积神经网络进行跨语料库的情感识别。此外,针对单一情感特征识别率不高的问题,提出一种新的特征融合方法。首先利用 OpenSMILE提取浅层声学特征,然后利用图卷积神经网络提取深层特征。随着卷积层的不断深入,节点的特征信息被传递给其他节点,使得深层特征包含更明确的节点特征信息和更详细的语义信息,然后将浅层特征和深层特征进行特征融合。采用两组实验进行验证,第一组用eNTERFACE库训练测试Berlin库,识别率为59.4%,第二组用Berlin库训练测试eNTERFACE库,识别率为36.1%。实验结果高于基线系统和参考文献中最优的研究成果,证明本文提出方法的有效性。
英文摘要:
      In reality, the speech data for speech emotion recognition are often collected from different databases. Hence, the discrepancy between the feature spaces of both will reduce recognition rate. In response to it, this paper proposes a new composition method to represent the topological structure between the source and target databases, and graph convolutional network is applied to cross-corpus emotion recognition. Besides, aiming at the problem of low accuracy of single feature in emotion recognition, a novel feature fusion method is proposed. First extract the acoustic features by OpenSMILE and then extract deep features by graph convolutional neural network. With going deeper of convolutional layers, the feature information of nodes will sends to other nodes, making the deep features contains clearer feature ?information and more detailed semantic information. Then perform feature fusion of the shallow and deep features. Two classification experiments are carried out. Firstly, eNTERFACE corpus is for training and Berlin corpus is for testing, and the recognition rate is 59.375%. Secondly, Berlin corpus is for training and eNTERFACE corpus is for testing, and the recognition rate is 36.111%. The experimental results are higher than the best research results in the baseline system and references, which proves the effectiveness of the method proposed in this paper.
  HTML  查看/发表评论

Copyright @2010-2015《数据采集与处理》编辑部

地址:南京市御道街29号        邮编:210016

电话:025-84892742      传真:025-84892742       E-mail:sjcj@nuaa.edu.cn

您是本站第2826599位访问者 本站今日一共被访问842

技术支持:北京勤云科技发展有限公司