Kevin’s research focus is on applications of rough path theory to landmark based human action detection, recognition and evaluation and gesture recognition. He is in particular interested in applications to sports data and human computer
interfaces, but his work has also included a project on monitoring safety on public escalators using CCTV. Here extracting landmark data from the video data using pose estimation has the crucial benefit of preserving the privacy
of people shown in the video.
Kevin's background is in Mathematics, with Computer Science as a minor. His research interests also include computer vision and applications of Functional Analysis, in particular Banach space theory, to the mathematical foundations
of machine learning.