Pro-Vice Chancellor for Learning, Teaching and Information
Areas of Expertise: Image Processing, Pattern Analysis
John Robinson is a Professor in the Department of Electronics and Pro-Vice Chancellor for Learning, Teaching and Information at the University of York. He has degrees in Mathematics (BSc), Telecommunications (MSc), Electronic Systems Engineering (PhD) and Humanities (MPhil). He has five years of industry experience in the UK and Canada, and twenty years as a University teacher, working at the University of Waterloo, Memorial University of Newfoundland and the University of York. He has published over 100 papers in the areas of image and video processing, pattern recognition and augmented reality. His hardware and software designs have appeared in products from Standard Telephones and Cables, British Telecom, Nortel, Bell Canada, Worldlinx Communications, Future Labs Inc., Intelligent Compression Technologies and others. He is active in University/Industry initiatives: he was a member of the management team for the Creation network; member of the management board of York Electronics Centre; co-founder and member of steering group for the Centre for Usable Home Technology (CUHTec) and former Director of the Higher York Creative Technologies Centre (CTC). He has served on various management and programme committees and consulted for UK and Canadian industry, Canadian provincial government and a major US law firm.
John usually teaches software-based subjects like Data Structures and Algorithms and modules in the Digital Media Systems stream. He has run the first year design competition since 2011.
John's current major research directions are in human face analysis, projector-camera systems and picture coding. For more on these see the Visual Systems Lab webpages.
A current focus is Face Description. We use a demographic/expression/landmark analysis method to describe pictures of people. For an example showing the processing rate achieved on a HiGrade Hotino C2200 laptop, see this video which illustrates processing of each frame independently (no temporal filtering of the results) using simple labels to show the estimations.
The method is based on conditional density estimation from mixture models.
We are working with industrial partners on several applications of the technology.
John currently sits on the executive committee of Professors and Heads of Department in Electrical and Electronic Engineering (PHEE). He is also involved in various external bodies in conjuction with his role as PVC Learning, Teaching and Information.