SubmissionNumber#=%=#16
FinalPaperTitle#=%=#Enhancing Stress Detection on Social Media Through Multi-Modal Fusion of Text and Synthesized Visuals
ShortPaperTitle#=%=#
NumberOfPages#=%=#10
CopyrightSigned#=%=#Efstathia Soufleri
JobTitle#==#Postdoctoral Researcher
Organization#==#Archimedes Unit, Athena Research Center
6 Artemidos and Epidaurou Str.
Marousi
Attiki 15125
Greece
Abstract#==#Social media platforms generate an enormous volume of multi-modal data, yet stress detection research has predominantly relied on text-based analysis. In this work, we propose a novel framework that integrates textual content with synthesized visual cues to enhance stress detection. Using the generative model DALL·E, we synthesize images from social media posts, which are then fused with text through the multi-modal capabilities of a pre-trained CLIP model. Our approach is evaluated on the Dreaddit dataset, where a classifier trained on frozen CLIP features achieves 94.90% accuracy, and full fine-tuning further improves performance to 98.41%. These results underscore the integration of synthesized visuals with textual data not only enhances stress detection but also offers a robust method over traditional text-only methods, paving the way for innovative approaches in mental health monitoring and social media analytics.
Author{1}{Firstname}#=%=#Efstathia
Author{1}{Lastname}#=%=#Soufleri
Author{1}{Username}#=%=#esoufleri
Author{1}{Email}#=%=#e.soufleri@athenarc.gr
Author{1}{Affiliation}#=%=#Athena RC
Author{2}{Firstname}#=%=#Sophia
Author{2}{Lastname}#=%=#Ananiadou
Author{2}{Username}#=%=#effie
Author{2}{Email}#=%=#sophia.nactem@gmail.com
Author{2}{Affiliation}#=%=#University of Manchester

==========
èéáğö