Depression Detection in Social Media Using Emotional Intelligence

Abstract

This project introduces a novel approach to detecting depression using multimodal deep learning, integrating convolutional neural networks (CNNs) for image analysis and long short-term memory (LSTM) networks for textual understanding. Implemented in MATLAB, the system leverages emotional intelligence from social media content to capture nuanced indicators of depression. By analyzing both images and text, the model aims to provide a comprehensive understanding of user’s emotional states, offering a promising avenue for early intervention and support. Two methods are employed: live face-based stress level detection and text-based stress level detection. This approach addresses the limitations of traditional detection methods by harnessing the rich emotional cues present in social media, thereby contributing to the mitigation of depression's escalating burden on individuals and society.