语音技术一电声技赢
oiceTechnologyuaioEngineering
文献引用格式：缪悦.基于自注意力机制的语音情感识别技术优化[J].电声技术，2025，49(1)：84-86.
MIAOY.Optimizationofspeechemotionrecognitiontechnologybasedonself-attentionmechanism[Jj.AudioEngineering,
2025,49(1):84-86.
中图分类号：TN912.34文献标识码：AD0l:10.16311/j.audioe.2025.01.025
基于自注意力机制的语音情感识别技术优化
缪悦
（云南广播电视台，云南昆明650500)
摘要：语音情感识别是通过分析和处理语音信号，识别出说话者的情感状态。随着人工智能技术的发展，自注意力机制逐渐
应用于语音情感识别领域。基于自注意力机制优化现有语音情感识别技术，提出一种自注意力机制方法，以提高识别精度和
健壮性。最后，通过实验验证了所提模型在多个公共数据集上的性能，并与现有主流方法进行了对比。实验结果表明，基于
自注意力机制的优化模型在情感识别的准确率、精确度、召回率等指标上均有显著提升。
关键词：自注意力机制；语音情感识别；人工智能
OptimizationofSpeechEmotionRecognitionTechnologyBasedonSelf-AttentionMechanism
MIAOYue
(YunnanMediaGroup,Kunming650500,China)
Abstract:Speechemotionrecognitionistoidentifythespeakersemotionalstatebyanalyzingandprocessingspeechsignals.With
thedevelopmentofartificialintelligencetechnology,self-attentionmechanismisgraduallyappliedtothefieldofspeechemotion
recognition.Basedontheself-attentionmechanism,theexistingspeechemotionrecognitiontechnologyisoptimized,andaself-
attentionmechanismmethodisproposedtoimprovetherecognitionaccuracyandrobustness.Finally,theperformanceoftheproposed
modelonseveralpublicdatasetsisverifiedbyexperimentsandcomparedwiththeexistingmainstreammethods.Theexperimental
resultsshowthattheoptimizationmodelbasedonself-attentionmechanismhassignificantlyimprovedtheaccuracy,precisionand
recallrateofemotionrecognition.