Climate and Weather: Inspecting Depression Detection via Emotion Recognition

TL;DR

This paper explores using emotion recognition features to improve automatic depression detection by leveraging multimodal data, demonstrating enhanced performance and stability on the DAIC-WOZ dataset.

Contribution

It introduces a novel approach of transferring emotion recognition knowledge to depression detection and fusing emotion, audio, and text modalities for better accuracy.

Findings

Improved depression detection performance on DAIC-WOZ.

Enhanced training stability with emotion transfer.

Insights into emotion expression in depressed individuals.

Abstract

Automatic depression detection has attracted increasing amount of attention but remains a challenging task. Psychological research suggests that depressive mood is closely related with emotion expression and perception, which motivates the investigation of whether knowledge of emotion recognition can be transferred for depression detection. This paper uses pretrained features extracted from the emotion recognition model for depression detection, further fuses emotion modality with audio and text to form multimodal depression detection. The proposed emotion transfer improves depression detection performance on DAIC-WOZ as well as increases the training stability. The analysis of how the emotion expressed by depressed individuals is further perceived provides clues for further understanding of the relationship between depression and emotion.