Abstract

The last decade has seen a revolution in the theory and application of computer vision. I will begin with a brief review of some of the fundamentals with a few examples of the reconstruction, registration and recognition of three-dimensional objects from uncalibrated RGB images and their translation into novel commercial applications.

I will then introduce two real-time deep learning systems that fully exploit geometry, compute model uncertainty and require very small amounts of labelled data. In particular, we will look at two practical examples we have recently developed of computer vision algorithms running in real-time on a mobile phone camera to recover 3D human body shape and to transcribe the readings on a hand-held glucose meter automatically and without using wireless connectivity.

Some references: Papers: https://mi.eng.cam.ac.uk/~cipolla/publications_selected.htm

Videos: https://www.youtube.com/user/ComputerVisionVideos