David Martin
Founder , 360 Cities

Website

My research interests are in image processing and machine vision. I am interested in the fundamental problem of how we parse an image (just a 2D grid of numbers!) into a coherent and accurate perception of surfaces in three dimensions. To this end, I’m interested in the low-level image features that provide the information to get started, as well as the higher level algorithms that stitch these features into visible surfaces, perceive shape, reason about occlusion, and perform amodal completion (seeing things behind other things). Apart from general image understanding, example applications are human body pose estimation, tracking, video coding, human-computer interaction, stereo vision, and omnidirectional imaging.