[1]胡婷婷,沈凌洁,冯亚琴,等.语音与文本情感识别中愤怒与开心误判分[J].计算机技术与发展,2018,28(11):124-127.[doi:10.3969/j.issn.1673-629X.2018.11.028]
　HU Ting-ting,SHEN Ling-jie,FENG Ya-qin,et al.Research on Anger and Happy Misclassification in Speech and Text Emotion Recognition[J].,2018,28(11):124-127.[doi:10.3969/j.issn.1673-629X.2018.11.028]
点击复制

语音与文本情感识别中愤怒与开心误判分()
分享到：

《计算机技术与发展》[ISSN:1006-6977/CN:61-1281/TN]

卷:
28
期数:
2018年11期
页码:
124-127
栏目:
智能、算法、系统工程
出版日期:
2018-11-10

文章信息/Info

Title:
Research on Anger and Happy Misclassification in Speech and Text Emotion Recognition
文章编号:
1673-629X(2018)11-0124-04
作者:
胡婷婷;  沈凌洁;  冯亚琴;  王蔚
南京师范大学 教育科学学院机器学习与认知实验室,江苏 南京,210097
Author(s):
HU Ting-ting; SHEN Ling-jie; FENG Ya-qin; WANG Wei
Lab of Machine Learning and Cognition of School of Education Science,Nanjing Normal University, Nanjing 210097,China
关键词:
情感识别;  声学特征;  文本特征;  混淆矩
Keywords:
emotion recognition; acoustic features; lexical features; confusion matrix
分类号:
TP18
DOI:
10.3969/j.issn.1673-629X.2018.11.028
文献标志码:
A
摘要:
在语音情感识别的研究中发现,愤怒与开心之间通过语音信息较难区分,文中将结合文本信息对这两种情感进行区分,以提高其识别率.使用IEMOCAP与SAVEE数据集中语音数据提取声学特征,分别使用卷积神经网络与支持向量机训练分类器模型,对中性、愤怒、开心、悲伤四类情感进行识别,对情感之间误判情况以及识别准确率进行分析,验证了语音通道信息对于愤怒与开心容易产生误判的结论.为解决此问题,加入文本信息,训练文本识别模型,有效解决了愤怒与开心的误判情况.同时发现两通道信息对情感识别的不同影响,在声音中包含更多利于识别愤怒和悲伤情感的信息;在文本中包含更多利于识别中性与开心情感的信息.声音情感识别中,愤怒/开心之间易误判,愤怒/悲伤之间易区分.文本情感识别中,愤怒/开心之间易区分,愤怒/悲伤之间易误判.声学与文本特征融合后,情感识别准确率相比单一通道明显提高,两通道信息对于情感识别具有互补作用.
Abstract:
In research on speech emotion recognition,it is easy to confuse anger with happy. We will combine text information to distinguish these two emotions to improve their recognition rate. The speeches of IEMOCAP and SAVEE data sets are used to extract acoustic features. Support vector machine (SVM) and convolution neural network (CNN) are used to recognize neutral,angry,happy,sad four types of emotion. Recognition accuracy of four emotions and misclassification are analyzed and it is concluded that the speech information can make misjudgment of anger and happy. To solve this problem,we add text contents to train text recognition model,resolving the misjudgment of anger and happy effectively. At the same time,we find the differences between two channels information in the emotion recognition. The features of acoustic have high accuracy of anger and sadness,low accuracy of neutral and happy,while text features have high accuracy of neutral and happy and low accuracy of anger and sadness. The anger and happy are easily misjudged in acoustic model, while they are easy to distinguish in text model. The anger and sadness are easily misjudged in text model,while they are easy to distinguish in acoustic model. After the fusion of the acoustic and text features,the accuracy is significantly improved compared to the single model,and the two models information have a complementary effect in the emotion recognition.

相似文献/References:

[1]韩志艳,伦淑娴,王健.基于遗传小波神经网络的语音情感识别[J].计算机技术与发展,2013,(01):75.
　HAN Zhi-yan,LUN Shu-xian,WANG Jian.Speech Emotion Recognition Based on Genetic Wavelet Neural Network[J].,2013,(11):75.
[2]韩志艳,王健. 多模式情感识别特征参数融合算法研究[J].计算机技术与发展,2016,26(05):27.
　HAN Zhi-yan,WANG Jian. Research on Feature Fusion Algorithm for Multimodal Emotion Recognition[J].,2016,26(11):27.
[3]李姗,徐珑婷. 基于语谱图提取瓶颈特征的情感识别算法研究[J].计算机技术与发展,2017,27(05):82.
　LI Shan,XU Long-ting. Research on Emotion Recognition Algorithm Based on Spectrogram Feature Extraction of Bottleneck Feature[J].,2017,27(11):82.
[4]李才隆,叶宁,黄海平,等.基于递归定量分析的生理信号情感识别[J].计算机技术与发展,2018,28(11):94.[doi:10.3969/ j. issn.1673-629X.2018.11.021]
　LI Cai-long,YE Ning,HUANG Hai-ping,et al.Physiological Signal Emotion Recognition Based on Recursive Quantitative Analysis[J].,2018,28(11):94.[doi:10.3969/ j. issn.1673-629X.2018.11.021]
[5]李田港,叶 硕,叶光明,等.基于集成学习的语音情感识别算法研究[J].计算机技术与发展,2020,30(06):82.[doi:10. 3969 / j. issn. 1673-629X. 2020. 06. 016]
　LI Tian-gang,YE Shuo,YE Guang-ming,et al.Research on Speech Emotion Recognition Algorithm Based on Ensemble Learning[J].,2020,30(11):82.[doi:10. 3969 / j. issn. 1673-629X. 2020. 06. 016]
[6]陶小梅*,陈心怡.在线学习环境中基于眼动特征情感识别研究[J].计算机技术与发展,2021,31(03):186.[doi:10. 3969 / j. issn. 1673-629X. 2021. 03. 032]
　TAO Xiao-mei *,CHEN Xin-yi.Research on Emotion Recognition Based on Eye Movement Features in e-Learning Environment[J].,2021,31(11):186.[doi:10. 3969 / j. issn. 1673-629X. 2021. 03. 032]
[7]游 兰,曾 晗,韩凡宇,等.基于 BERT-BiGRU 集成学习的情感语义识别[J].计算机技术与发展,2023,33(05):159.[doi:10. 3969 / j. issn. 1673-629X. 2023. 05. 024]
　YOU Lan,ZENG Han,HAN Fan-yu,et al.Sentiment Semantic Recognition Based on BERT-BiGRU Ensemble Learning[J].,2023,33(11):159.[doi:10. 3969 / j. issn. 1673-629X. 2023. 05. 024]

更新日期/Last Update: 2018-11-10