Emotion-Aware ResNet50V2: Enhancing Mental Health Detection through Facial Expression Analysis

Mental health conditions, such as anxiety, depression, and stress, significantly impact individuals across diverse demographics. Despite advances in awareness, many cases remain undiagnosed or untreated. This study introduces Emo-Res50V2, a customized ResNet50V2 architecture, to detect facial emotions accurately using the FER2013 dataset. By incorporating an emotion-aware classifier, our model achieves 90.03% accuracy. We correlate detected emotions with mental health conditions through survey data, providing a comprehensive tool for emotional and psychological assessment. Our approach demonstrates robustness against noisy data, outperforming state-of-the-art techniques. This research highlights the potential of deep learning in advancing mental health detection, facilitating early diagnosis, and personalized treatment planning. The code is available at https://github.com/Myself-Rohit-Dey/Emo-Res50V2