DHRITI: A RESPONSIVE AFFECTIVE MUSIC COMPANION USING HOG AND PLUTCHIK’S EMOTION MODEL FOR HUMAN EMOTIONAL WELLBEING

Current smart assistants, despite their widespread adoption, remain limited by their inability to interpret and adapt to human emotions, diminishing their potential as emotionally intelligent systems (EIS). To address this gap, we propose Dhriti, a Human-Centered Affective System that integrates Histogram of Oriented Gradients (HOG)-based facial emotion recognition with adaptive music therapy to create a Responsive Affective System (RAS). Dhriti functions as an Emotional Wellbeing Platform (EWP), leveraging computer vision to analyze facial expressions and physiological cues, enabling real-time detection of emotions such as happy (joy), sad, angry, and surprise. This data drives an Adaptive Emotional Interface (AEI), which generates responses through context-aware interventions, including dynamically curated music therapy, empathetic dialogue, and stress-alleviation strategies. By synthesizing HOG’s robustness in feature extraction with affective computing principles, Dhriti transcends conventional rule-based interactions, evolving into a Responsive Emotional Companion. The system’s architecture prioritizes emotional congruence, ensuring responses align with the user’s psychological state—for instance, offering calming melodies during detected stress or uplifting tracks to reinforce positive affect. This paper reviews recent advancements in RAS frameworks, contextualizes Dhriti’s methodology within the EIS paradigm, and evaluates its efficacy in enhancing emotional wellbeing through empirical user studies. Results demonstrate its capacity to foster meaningful human-AI rapport, positioning Dhriti as a scalable model for next-generation emotionally intelligent interfaces.