About VSL

Vision Scientific Ltd, is an Open University research spinoff company founded by Professors Jeff Johnson and Phil Picton in 1989.

Our mission is to pursue scientific research and education through applications in the public and private sectors.

The company was born out of an academic project in the nineteen eighties with British Telecom Research Laboratories investigating the use of neural networks through the case study of recognising eyes in faces. We took a radically different approach to this problem using an approach based on discrete algebra rather than continuous analytic mathematics. For us the fundmental idea was that the greyscales of  digital images determine relations between neigbouring pixels. In particular we experimented with classifying pixels according to whether their greyscales were lighter or darker than each of their four-neighbours, where four brighter/darker binary relations for four neighbours gives sixteen possible classes (ignoring greyscale equality). We were astonished when colour coded displays of the pixels showed that they formed coherent polygons, leading to our algebraic concept of gradient polygons. Vision Scientifc Ltd was formed to develop these ideas in the context of real world applications where the measure of success is solving clients' problems and adding value.

Our first research contracts were with Unilever Research in the early nineteen ninties. At that time Unilever was researching hair growth by measuring changes in the length and diameter of individual hairs on scalp sites about 1 cm² over an interval of three days. Typically a hair has diameter 40 - 100 microns and making accurate measurements of diameter and growth is a challenging problem. Unilever had developed a method involving 35 mm photographs of the scalp enlarged to A4 prints. Operators then measured the lengths and diameters using a graphics tablet and stylus. Needless to say this method gave poor results. A short period of research into this problem led to our concept of gradient runs which are parameter-free objects enabling very robust feature detection in noisy and low-contrast images. In a series of contracts we developed a sub-pixel accurate method of measuring hair length and diameter which we built into complete turnkey measurement systems which we supplied and supported over a period of about five years. It was a pleasure to work with the Unilever scientists and we gained a lot of knowledge and experience in this collaboration.

Following this we have worked with various organisations including AOARD (the Asian Office of Aerospace Research and Development of the US Department of Defense) and A2iA (Artificial Intelligence and Image Analysis based in Paris), had various collaborations with the European Commission, and developed educatonal image processing software with the Open University in the UK.

Since 2006 the activities of VSL have focussed on fundamental research developing our ideas. This research culminated in the publication of the book Hypernetworks in the Science of Complex Systems published by Imperial College Press in London in 2014. Having established a strong research base, VSL is entering a new phase of applied research and commerical activity in three areas:

• machine vision: using new concepts recently developed, VSL is now actively seeking cutting-edge challenges in machine vision. We particularly look for difficult problems in machine vision that have not been solved by other means. We will give a confidential appraisal of any problem at no cost, and we work on the basis of 'no added-value solution - no fee'.

• education: for the past four years, working with the Open University and the European Commission, we have been researching scalable education for applications in the public and private sectors. We have created an Intelligent Peer Marking Platform to support low-cost no-cost education by providing high quality student peer assessment based on detecting and discriminating good and bad markers using hypernetwork principles. Available at no cost to the research community, this platform is available for low-cost high added-value applications in business and commerce. We are currently working on the European Erasmus DA.RE project: Data Science Pathways to reimagine eduction.
  

• transportation modelling: we have a strong background in research to provide methods for modelling large multilevel transportation systems, integrating traffic dynamics from the level of individual road intersections to cities, regions and nations. We are working with academic and commercial partners towards major applications in our local city of Milton Keynes, London and other cities worldwide.

• Global Systems Science. We are exploring the possibility of developing new software to support 'joined up' policy at all levels.