Aakash Kaku

Papers

Total Citations

H-Index

About

Aakash Kaku is a researcher at the forefront of machine learning for healthcare and human motion analysis. His primary research focuses on developing high-temporal-resolution action identification systems, with a particular emphasis on fine-grained movements critical for clinical applications. Kaku’s major contribution is the creation of **StrokeRehab**, a benchmark dataset for sub-second action identification, which addresses a key gap in the field: most existing models identify coarse actions (e.g., running) but fail to capture the rapid, subtle motions essential for rehabilitation monitoring. His work on sequence-to-sequence modeling for action identification further advances the ability to parse continuous, high-speed kinematic and video data. While his most-cited papers have accrued 5 and 2 citations respectively, their impact lies in pioneering a new direction for smart health and robotics—enabling systems to detect and classify movements at a temporal resolution previously unattainable. This research has direct implications for automated stroke rehabilitation assessment, where precise, real-time feedback can significantly improve patient outcomes. Kaku’s work is foundational for researchers aiming to bridge the gap between coarse activity recognition and the nuanced, sub-second actions that define clinical and therapeutic contexts.