Gita-Prerna: An AI-Based Emotion-Aware Guidance System Using Facial Emotion Recognition and NLP-Driven Shloka Recommendation

Abstract

Human emotions strongly influence mental well-being, decision-making, and social interaction. With the growing dependence on digital platforms, intelligent emotional support systems have gained significant attention in recent years. This paper presents Gita-Prerna, an AI-driven emotion-aware guidance system integrating Facial Emotion Recognition (FER), deep learning, and Natural Language Processing (NLP) to pro-vide personalized motivational guidance using Bhagavad Gita shlokas and inspirational slogans. The proposed framework captures real-time facial expressions through video input, identifies dominant emotions using CNN-based deep learning techniques, and retrieves contextually relevant motivational content using semantic recommendation mechanisms. The FER module was trained using the FER2013 dataset and achieved an overall emotion classification accuracy of 92.4%. The proposed system integrates culturally meaningful guidance with intelligent recommendation techniques to create a more personalized emotional support experience suitable for emotional wellness applications and intelligent virtual assistants

Introduction

Human emotions strongly affect behavior, communication, productivity, and mental well-being. People often experience stress, anxiety, frustration, confusion, and sadness due to academic, professional, and social pressures. Most existing emotional wellness applications rely on manual user input and cannot accurately identify a person's true emotional state.

The proposed framework, Gita-Prerna, uses Artificial Intelligence, computer vision, and Natural Language Processing (NLP) to provide personalized emotional guidance. It employs Facial Emotion Recognition (FER) to detect emotions from facial expressions, which are rich indicators of psychological states. Deep learning models classify emotions such as happiness, sadness, anger, fear, and neutrality, while an NLP-based recommendation system delivers relevant motivational content and Bhagavad Gita teachings.

Current systems have several limitations, including dependence on text-based sentiment analysis, lack of real-time emotion detection, static recommendations, and poor integration between emotion analysis and recommendation engines. These shortcomings reduce personalization and contextual understanding.

To address these issues, Gita-Prerna integrates real-time facial emotion recognition with semantic recommendation techniques. The system captures video input, preprocesses facial images, and uses a deep learning model trained on the FER2013 dataset to identify emotions. The detected emotional state is then passed to a recommendation engine that uses BERT/SBERT-based semantic similarity to retrieve appropriate Bhagavad Gita shlokas, explanations, motivational slogans, and optional audio content.

The architecture consists of a user interface, API layer, emotion detection module, and recommendation engine. The workflow includes video capture, facial feature extraction, emotion classification, and personalized content generation. Additional features include emotion history tracking, favorite shloka saving, audio playback, and adaptive recommendations.

Conclusion

This paper presented Gita-Prerna, an AI-based emotional guidance framework integrating Facial Emotion Recognition, deep learning, and NLP-driven recommendation systems. The proposed system analyzes facial expressions through video input and provides personalized Bhagavad Gita shlokas and motivational guidance aligned with the detected emotional state. The framework demonstrates how emotionally intelligent AI systems can support mental wellness through adaptive, context-aware, and culturally enriched recommendations. Future enhancements may include multilingual support, voice-based emotion recognition, reinforcement learning-based personalization, mobile application deployment, and integration with wearable healthcare systems.

References

[1] P. Ekman and W. V. Friesen, “Facial Action Coding System,” Consulting Psychologists Press, 1978. [2] I. Goodfellow et al., “Challenges in Representation Learning: A Report on Three Machine Learning Contests,” Neural Networks, vol. 64, pp. 59–63, 2015. [3] A. Mollahosseini, D. Chan, and M. H. Mahoor, “Going Deeper in Facial Expression Recognition Using Deep Neural Networks,” IEEE Winter Conference on Applications of Computer Vision, 2016. [4] J. Devlin et al., “BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding,” NAACL-HLT, 2019. [5] N. Reimers and I. Gurevych, “Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks,” EMNLP, 2019. [6] R. Picard, Affective Computing, MIT Press, 1997. [7] The Bhagavad Gita, Translated by Eknath Easwaran, Nilgiri Press, 2007. [8] M. Abadi et al., “TensorFlow: Large-Scale Machine Learning on Het-erogeneous Systems,” 2016. [9] G. Bradski, “The OpenCV Library,” Dr. Dobb’s Journal of Software Tools, 2000.

Copyright

Copyright © 2026 Para Upendar, CH. Harshini, CH. Rahul, D. Samiksha, D. Cheruthan. 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.