About

I build vision systems that understand interaction — how hands, objects, and scenes move and relate in 3D. The methods have ranged from probabilistic tracking and particle filters through deep learning to differentiable rendering and neural scene representation, but the goal is the same: make a system that explains what it sees.

Over the years this has meant work across computer vision, machine learning, robotics, and systems — in academia, in EU projects, and in industry.

A more detailed record is in my CV.

I did my degrees and post-doc at the University of Crete and FORTH, working with Prof. Antonis Argyros.

This is an early (2014) PoC which I still find inspiring. It was computed through this work. The head, hands, peripherals, cup and desk are participating in 3D tracking, jointly. Input was RGBD.