【摘要】 针对抑郁症的确诊过程繁琐，且依赖于医师的主观判断和经验等问题，本研究利用计算机算法提出了一种基于面部运动单元的抑郁症检测系统。该方法通过扩大人脸图像识别范围，并引入专家先验知识，设置面部运动单元分割规则将人脸检测模型与人脸关键点检测模型的输出结果编码到抑郁症检测模型中；最后，根据获取到的人脸图像中的关键点对每张人脸图像进行区域划分，实现在更细粒度上对人脸的局部区域进行面部运动单元的识别，提高抑郁症检测的精度和准确性。本研究可为基于面部运动单元的抑郁症检测提供一种新思路，具有重要研究意义。更多还原

【Abstract】 In order to solve the problem that the depression diagnosis in medical practice is complicated and depends on the subjective judgment and experience accumulation of doctors, we proposed a depression detection system based on facial action unit by using computer algorithm, which expanded the range of facial image recognition and introduced expert prior knowledge. The output results of the face detection model and the key point detection model were encoded into the depression detection model by setting the segmentation rules of the facial action unit. Finally, each face image was divided into regions, which could achieve facial action unit recognition of local regions of the face at a finer granularity, thereby achieving higher accuracy and improving the accuracy of depression detection. This study can provide a new idea for depression detection based on facial action unit, which has important research significance.更多还原