Ashwini Sawant

Papers

Total Citations

H-Index

About

Ashwini Sawant is a researcher advancing the frontier of human-computer interaction through multimodal affect recognition. Her work centers on developing deep learning frameworks that enable machines to perceive and interpret human emotional states from real-world audio-visual data. Sawant’s key contribution is the Multimodal Diverse Spatio-Temporal Network (MDSTN), a low-complexity architecture designed for video-based affect recognition. This model addresses the limitations of unimodal approaches—such as relying solely on speech or visual cues—by fusing multiple modalities to capture the nuanced, dynamic expression of emotions. Her research has direct implications for critical applications, including stress and depression assessment, Alzheimer’s disease detection, healthcare, and assistive robotics. Though early in her career, with her most-cited papers from 2024 and 2025 each garnering 2 citations, Sawant’s work is already recognized for its practical potential in real-world, unconstrained environments. By tackling the challenge of robust, multimodal perception, she is laying the groundwork for more empathetic and responsive interactive systems, making her a promising voice in affective computing and human-robot interaction.