Manasvita : AI-Powered Multimodal Mental Wellness Platform

Abstract

Mental health issues such as stress, anxiety, and depression have become increasingly prevalent in today’s fast-paced world. Early detection and timely intervention can significantly improve mental well-being. This project presents a Manasvita : AI-Powered Multimodal Mental Wellness Platformintegrating advanced machine learning techniques and AI-driven solutions to assist users in understanding and managing their mental health. The system incorporates user authentication, enabling secure access to personalized assessments and services. A doctor appointment module allows users to schedule consultations with mental health professionals. The platform utilizes Convolutional Neural Networks (CNNs) to analyze facial expressions and predict mental health conditions based on the FER2013 dataset. Additionally, a Random Forest classifier assesses stress levels using a structured dataset. An AI-powered chatbot, leveraging the Gemini AI API, provides users with immediate mental health-related support and guidance. To encourage positive behavioral changes, the platform includes a Task and Reward system, where doctors assign therapeutic tasks, and users earn incentives such as discounts or coupons upon completion. By integrating machine learning, artificial intelligence, and user engagement strategies, this project aims to provide an accessible, technology-driven solution for mental health monitoring and support. The proposed system enhances self-awareness, promotes timely intervention, and bridges the gap between users and professional healthcare services

Introduction

Mental health issues like stress, anxiety, and depression are increasingly common, yet many struggle to access timely support. The project presents Manasvita, an AI-powered multimodal mental wellness platform designed to help users evaluate their mental health and connect with professionals. It integrates several AI and machine learning components:

• Facial expression-based detection using CNN trained on the FER2013 dataset to recognize emotional states.

• Stress, anxiety, and depression prediction via a Random Forest model analyzing behavioral and physiological inputs.

• An AI chatbot powered by Gemini AI API to provide real-time mental health guidance and answer queries.

• A doctor appointment system for scheduling professional consultations.

• A task and reward system to motivate users through gamification, encouraging adherence to mental wellness activities.

Built on a web platform using the MERN stack with Flask APIs for model deployment, Manasvita aims to improve accessibility, early detection, and proactive support for mental health.

The literature review highlights the advances and limitations in AI mental health applications, emphasizing the need for personalization, real-time monitoring, ethical considerations, and integration with professional healthcare. The methodology involves data collection, preprocessing, model training, and system integration with a focus on privacy and security compliance.

Conclusion

The development of this mental health web application represents a significant step toward integrating artificial intelligence into mental healthcare, providing users with accessible and technology-driven solutions for mental health assessment. With an increasing prevalence of mental health disorders worldwide, there is a pressing need for scalable and efficient tools that can assist individuals in understanding their mental state and seeking timely interventions [1]. This project successfully addresses this need by incorporating various AI-driven features, including facial expression recognition using CNN, stress analysis through Random Forest, an AI chatbot, and a doctor appointment system. The CNN-based facial emotion recognition system plays a crucial role in identifying a user’s emotional state based on visual cues, allowing for an initial mental health assessment [3]. Since facial expressions often provide deep insights into a person’s psychological condition, this feature enhances the accuracy and reliability of early mental health detection. However, real-world challenges such as variations in lighting conditions, diverse facial expressions, and dataset biases still need to be addressed for improved generalization and accuracy. Similarly, the Random Forest-based stress, anxiety, and depression prediction model utilizes structured data to analyze behavioral patterns and assess the likelihood of mental health conditions [4]. By combining machine learning techniques with mental health indicators, the system provides users with a data-driven understanding of their well-being. Despite the promising results, continuous model refinement with larger and more diverse datasets is necessary to enhance the reliability and accuracy of predictions. The AI-powered chatbot, implemented using the Gemini AI API, further contributes to the platform by offering real-time, conversational support to users experiencing stress, anxiety, or depression [5]. This chatbot helps bridge the gap between professional mental health assistance and self-help resources, providing users with immediate guidance and coping strategies. While AI chatbots significantly improve accessibility, they still face limitations in understanding deep emotional contexts and providing personalized therapy. Another critical component of this project is the doctor appointment system, which connects users with mental health professionals. This feature ensures that individuals who require expert consultation can schedule appointments efficiently, reducing the barriers to professional help. Such integrations align with modern telehealth approaches, promoting mental health awareness and care accessibility [7].

References

[1] World Health Organization. (2021). Mental health: Strengthening our response. Retrieved from https://www.who.int/news-room/fact-sheets/detail/mental-health-strengthening-our-response [2] K. K. Aggarwal et al., \"The Role of Technology in Mental Health Care\", International Journal of Mental Health Systems, vol. 12, no. 3, pp. 45-60, 2020. [3] M. Sambare, \"FER2013 Dataset for Facial Emotion Recognition\", Kaggle, 2013. Retrieved from https://www.kaggle.com/datasets/msambare/fer2013 [4] S. John, \"Stress Level Prediction Dataset\", Kaggle, 2021. Retrieved from https://www.kaggle.com/datasets/shijo96john/stress-level-prediction/data [5] Google AI, \"Gemini AI: The Future of Conversational AI\", Google Research, 2023. Retrieved from https://ai.googleblog.com/ [6] J. D. Smith and R. White, \"Gamification in Mental Health: The Impact of Rewards on User Engagement\", Journal of Digital Health Research, vol. 8, no. 2, pp. 112-128, 2022. [7] National Institute of Mental Health, \"Bridging the Gap: The Role of Digital Platforms in Mental Health\", 2021. Retrieved from https://www.nimh.nih.gov/ [8] R. Patel, S. Kumar, and L. Desai, “Predicting Anxiety and Depression using Machine Learning Models,” Journal of Psychological Data Science, vol. 5, no. 1, pp. 23-38, 2020. [9] L. Wang and H. Zhou, “CNN-based Emotion Recognition for Mental Health Monitoring,” IEEE Transactions on Affective Computing, vol. 10, no. 4, pp. 512-523, 2019. [10] A. Gupta, M. Kaur, and R. Sen, “AI Chatbots for Mental Health: A Review on Effectiveness in Therapy Sessions,” Journal of AI in Healthcare, vol. 9, no. 3, pp. 78-95, 2022. [11] B. Kim and Y. Lee, “Deep Learning for Mental Health Analysis using Social Media Data,” Journal of Computational Psychiatry, vol. 6, no. 2, pp. 34-50, 2021.

Copyright

Copyright © 2025 Ms. Anuradha Singh, Vasudev Jaiswal, Satyam Singh Gaur, Aditya Yadav, Sakshi Goyal. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.