Contacts: Professor Graham Finlayson, Dr. David Connah, School of Computing Sciences, UEA
Mentor: Dr. Mark Priest, Harrogate Partners

For some visual examples of the technology’s capabilities, see here

Professor Finlayson, founder of university spin-out Imsense which was successfully acquired in mid-2010, and his colleague David Connah, have developed a new approach to the problem of image fusion, which they hope will provide the seed for a new start-up company. Image Fusion is where multiple sources of information, for example visual photographs plus infrared measurements, are combined into a single image. This can make it easier to “see through” fog, or just provide a concise way to communicate multiple measurements,

Although false colour infrared pictures are now commonplace, these can often create a confusing output when combined with a visual photo. Typical problems include creating artefacts so that false objects are seen in the image, or losing important details from the original visual image. While people can be trained to interpret even quite confusing images, in situations when they are interacting with their surroundings in real-time it is crucially important that the image is not so confusing as to slow their reactions and responses. Artefacts can also cause a range of problems, since it is often quite difficult to distinguish between a real object and an artefact.

The new technology solves these problems in a unique patented way, using gradient domain processing. This results in images which are artefact-free and which also look very similar to the usual visual world. In particular the core visual information is not degraded by the image fusion process.

Image fusion is already used in a wide variety of situations, even with the limitations of existing methods. These include aerial and satellite imaging, where multiple data sources can be combined into a single visual image, medical imaging where the outputs of several different scanning techniques may be combined, and even robot vision. There are also a number of military applications, most of which need accurate real-time information.

The challenge for the i-Team is to investigate and recommend the best uses of this new technology outside of the military field. This is expected to include the wide range of current uses for image fusion, as well as any new uses that the i-Team uncovers. Some may be general uses for image fusion technology where this technology can outperform existing solutions, while other uses may be very specific to the way in which the new method works and visualises the information.