Speaker "Vijay Nadkarni" Details Back



Deep Learning-Based Human Activity Recognition for Physical Skills Improvement


The ability to recognize and analyze human activities is becoming increasingly important in a multitude of industries, including healthcare, sports, industrial operations and numerous others. The fusion of computer vision and physical sensing is key to this. Ranging from simple human activities such as patients performing rehab routines to complex ones such as the skills assessment of assembly line operators or technique of sports players, there now exists the ability to provide training and improvement of the physical skills of individuals. This presentation will cover the AI technologies at play here, including pose estimation and human action recognition (HAR), and provide an overview of the deep learning algorithms that are used. In addition, a summary of the challenges that are presently being researched will be presented.


Vijay Nadkarni is Chief Technology Officer of Movella. He is responsible for Movella’s technology strategy advancement and the alignment of the company’s business goals and strategic plan with its technological vision and roadmap. This involves embracing four key areas, namely AI, SaaS, motion analytics and computer vision. Vijay has over 25 years in technology leadership and management across a wide range of industries. Prior to Movella, he served as VP of Artificial Intelligence at Tech Mahindra where he headed its AI Practice across 10 diverse verticals. Before that, Vijay was VP of Artificial Intelligence at Visteon with ownership of AI across its product lines, notably autonomous driving and infotainment. A veteran of Silicon Valley, Vijay has co-founded multiple startups in AI, motion-analytics and cloud among which Veraz Networks - a VoIP company - had a NASDAQ IPO. Vijay has an MBA and MSEE, both from Northwestern University, and a BS in Electrical Engineering from IIT-Bombay. In his spare time, Vijay enjoys biking and is an avid AI hobbyist who loves writing code for challenging projects.