Bhava to Bytes: A Multimodal Fusion of Indian Culture and Emotion-Aware AI

Abstract

This work explores the relationship between deep emotional frameworks embedded in Indian heritage, tradition, and culture and multimodal affective computing. India offers a rich tapestry of emotional expression through Ayurveda, traditional dance, music, and spiritual practices that have their roots in philosophies like Navarasa and Bhava-Rasa. These age-old methods convey complex emotions through movement, song, and ritual. Through the integration of these cultural idioms with modern computational techniques.

Emotionally intelligent technology must be inclusive, culturally aware, and grounded in a range of expressive traditions. RASA-Net proposed a multimodal deep learning framework that blends India's traditional Rasa theory with state-of-the-art emotion recognition. By combining visual cues from Kathak and Bharatanatyam dance, vocal tonality guided by Sanskrit, text in the Indic language, and yogic physiological markers, RASA-Net offers culturally nuanced emotion classification. Comparative evaluations reveal significant advancements over largely Western-centric baselines, especially in the areas of complex emotional states and abrupt changes. Ethical guidelines for algorithmic accountability, data dignity, and cultural consent guarantee responsible deployment.