基于语音预训练模型的抑郁症识别研究  

A research on depression recognition based on voice pre-training model

作　　者：黄祥胜 廖义龙 张文劲 张莉[1] HUANG Xiangsheng;LIAO Yilong;ZHANG Wenjing;ZHANG Li(School of Biomedical Engineering,South-Central Minzu University,Wuhan 430074,P.R.China)

机构地区：[1]中南民族大学生物医学工程学院,武汉430074

出　　处：《生物医学工程学杂志》2024年第1期9-16,共8页Journal of Biomedical Engineering

基　　金：国家自然科学基金资助项目(81601586);中央高校基本科研业务费专项资金项目(CZZ21007,CZQ23031,CZQ23029)。

摘　　要：针对数量日益增多的抑郁症患者群体,本文提出一种通过语音信号有效识别抑郁症的人工智能方法,以提高诊断和治疗效率。首先,通过微调语音到特征向量模型2.0(wav2vec 2.0)的预训练模型对语音进行编码和上下文化,从而获得高质量的语音特征;然后,应用情感障碍分析的公用数据集——绿野仙踪忧虑分析访谈语料库(DAIC-WOZ)数据集验证上述模型。结果显示,在抑郁症识别的二分类任务中,该方法在精确率方面达到了93.96%、召回率达到了94.87%、F1分数达到了94.41%,总体分类准确率达到96.48%。在评估抑郁症严重程度的四分类任务中,精确率均达到92.59%及以上,召回率均达到92.89%及以上,F1分数均达到93.12%以上,总体分类准确率达到94.80%。基于上述结果证明,本文提出的方法在小样本情况下有效提升了分类的准确率,对于抑郁症的识别和严重程度评估效果良好。未来,该方法有望在抑郁症的诊断中起到辅助支持的作用。For the increasing number of patients with depression,this paper proposes an artificial intelligence method to effectively identify depression through voice signals,with the aim of improving the efficiency of diagnosis and treatment.Firstly,a pre-training model called wav2vec 2.0 is fine-tuned to encode and contextualize the speech,thereby obtaining high-quality voice features.This model is applied to the publicly available dataset-the distress analysis interview corpus-wizard of OZ(DAIC-WOZ).The results demonstrate a precision rate of 93.96%,a recall rate of 94.87%,and an F1 score of 94.41%for the binary classification task of depression recognition,resulting in an overall classification accuracy of 96.48%.For the four-class classification task evaluating the severity of depression,the precision rates are all above 92.59%,the recall rates are all above 92.89%,the F1 scores are all above 93.12%,and the overall classification accuracy is 94.80%.The research findings indicate that the proposed method effectively enhances classification accuracy in scenarios with limited data,exhibiting strong performance in depression identification and severity evaluation.In the future,this method has the potential to serve as a valuable supportive tool for depression diagnosis.

关 键 词：抑郁症识别 语音预训练模型 语音特征 

分 类 号：R749.4[医药卫生—神经病学与精神病学] TN912.3[医药卫生—临床医学] TP18[电子电信—通信与信息系统]