Enhancing Depression Detection on TikTok: A Multi-model Approach Leveraging ImageBand

Summary

Depression is a prevalent mental health condition characterized by persistent sadness and a lack of interest or pleasure in activities that were once rewarding or enjoyable. According to recent data from the World Health Organization (WHO), an estimated 3.8% of the global population suffers from depression. This includes 5% of adults, with a prevalence of 4% among men and 6% among women. The situation can be particularly challenging for individuals who lack access to treatments for mental disorders. Given the considerable public health burden of depression and the limitations of current detection methods, this project seeks to 1) identify quantitative indicators in TikTok posts that align with depressive symptom content from the PHQ-9 and SMFQ items; and 2) optimize a deep learning model with high accuracy to identify depressive symptoms more objectively and effectively.

Team

Yanfang Wu (Journalism and Media Management), Mitsunori Ogihara (Computer Science), Spencer Evans (Psychology)

Enhancing Depression Detection on TikTok: A Multi-model Approach Leveraging ImageBand