• Shaikh Mohd Ashfaque
• Saif H Shaikh
• Hemal Naik
• Krishna Handa
• Atharv Rasal

Speech Emotion Recognition (SER) is an interdisciplinary area that enhances machine understanding of human emotions through voice. It holds vital significance for fields such as healthcare, assistive technologies, intelligent virtual assistants, and affective computing. This paper introduces a comprehensive SER system leveraging deep learning and context-awareness using the Wav2Vec2 transformer model. Our solution uses raw speech input, avoids traditional feature engineering, and improves emotion classification accuracy by integrating speech transcription with sentiment analysis. Using the RAVDESS dataset, we trained and evaluated our model to classify eight emotions with real-time inference capability. Results show 78% overall accuracy with high performance for emotions like Calm, Disgust, and Surprise. This system showcases the future potential of context-enriched emotional AI systems.

Audio Classification, Context-Aware Analysis, Deep Learning, Human-Computer Interaction, Real-Time Systems, Speech Emotion Recognition, Transformer Models, Wav2Vec2.

"Multi-Modal Speech Emotion Recognition: Integrating Transformer Models and Contextual Analysis", International Journal of Emerging Technologies and Innovative Research (www.jetir.org), ISSN:2349-5162, Vol.12, Issue 4, page no.n347-n360, April-2025, Available :http://www.jetir.org/papers/JETIR2504D45.pdf

"Multi-Modal Speech Emotion Recognition: Integrating Transformer Models and Contextual Analysis", International Journal of Emerging Technologies and Innovative Research (www.jetir.org | UGC and issn Approved), ISSN:2349-5162, Vol.12, Issue 4, page no. ppn347-n360, April-2025, Available at : http://www.jetir.org/papers/JETIR2504D45.pdf