摘要
语音情感识别是具有重要应用价值和具有挑战性的研究问题。该文提出了一个旨在提高语言情绪识别准确性的研究。该研究使用效价-激活度二维情绪空间模型对英语语言情绪进行描述,使用基于高斯混合模型和支持向量机的分阶段方法来对英语发言者的情绪效价(积极与消极情绪)和情绪激活度(情感的强度)进行分类识别。最后通过实验验证了该算法能够较为正确地识别使用英语表达的不同情绪。
Speech emotion recognition is an important and challenging research question.This paper presents a study aimed at improving the accuracy of language emotion recognition.This study describes the English language sentiment using the two-dimensional emotional space model of potency-activity,and uses the staged approach based on Gaussian mixture model and support vector machine to measure the emotional valence(positive and negative emotions)and emotion Activation(emotional intensity)for classification and identification.Finally,experiments show that the algorithm can identify different emotions expressed in English more correctly.
作者
高成吉
GAO Chengji(Xi'an Aeronautical Polytechnic Institute, Xi'an 710089)
出处
《计算机与数字工程》
2019年第7期1622-1626,共5页
Computer & Digital Engineering
关键词
情绪识别
情绪效价
激活度
高斯混合模型
支持向量机
emotion recognition
emotional valence
arousal
gaussian mixture model
support vector machine