An Applied Study of Multi-modal Emotion Recognition to Assist in The Diagnosis of Depression

Abstract

Multi-modal emotion recognition significantly enhances the reliability and precision of detecting emotions by utilizing complementary information between different modalities and comprehensively analyzing physiological signals such as text, audio, and video. The rapid development of multi-modal emotion recognition technology has led to its wide application in the fields of medical diagnosis, educational feedback, and human-computer interaction, and has provided a new way for the objective diagnosis of depression. The aim of this paper is to review the latest progress of multi-modal affective computing for assisted depression diagnosis technology. The review will cover two aspects: application examples of multi-modal affect recognition in assisted depression diagnosis, future trends and research directions of multi-modal affect recognition. The paper aims to explore and analyze the application and efficacy of the three multi-modal affect recognition models proposed in the past three years for assisted depression diagnosis. The design principles and diagnostic effects of each model will be outlined. The current challenges and potential of their research will be explored. The performance of these models for improving diagnostic accuracy and assisting depression diagnosis will be evaluated and compared. The paper will discuss the current challenges of their research and their potential in practical applications., The paper will explore the contribution of multi-modal data fusion to improving diagnostic accuracy and look forward to future developments.