摘要
Abstract
This paper introduces a multimodal intelligent depression-assisted diagnosis system and its data privacy protection sub-project.The system combines multimodal data on facial expressions,speech and words to improve the accuracy and efficiency of depression diagnosis.In view of the privacy protection of the video image data set of depression patients,the project adopts OpenFace 2.2.0 for facial feature extraction and encryption processing,to ensure that the information needed for diagnosis is retained while preventing the exposure of specific facial features of patients.We designed a strict user application and authorization mechanism to ensure that access and use of datasets is limited to authorized users.In addition,it provides protection measures for users'self-processing data,and embedding APIs to ensure the security and convenience of data processing.In the future,we will continue to optimize the privacy protection mechanism,improve the security of the system and user trust,and provide more reliable support for intelligent diagnosis of depression.关键词
多模态/智能诊断/抑郁症/数据隐私保护/面部特征提取Key words
multimodal/intelligent diagnosis/depression/data privacy protection/facial feature extraction分类
信息技术与安全科学
