Image Analysis Group

Image Analysis is a long-established cross-functional group of scientific researchers, software engineers and clinical specialists. Now part of our AI Centre of Excellence, the teams in the Image Analysis group work on applied research and development of imaging algorithms using state-of-the-art deep learning as well as classical techniques. We develop algorithms for use across multiple clinical areas: neurology, cardiology, oncology and MSK. The algorithms we develop are delivered as software components. Our colleagues in Japan and US integrate them into product applications, ultimately deployed for use in medical sites around the world. As well as working collaboratively with colleagues in the global Canon Medical family, our team members also work with international clinical collaborators.

To ensure we are truly working on state-of-the-art research and development, continuous upskilling is key to our success. Our group members attend medical image analysis and machine learning courses and conferences, absorb knowledge from clinical collaborators and have an internal training strategy to support keeping abreast of new methods, ideas and technology.

Project Spotlight

Image of Embolisation Plan

The Embolisation Plan algorithm tracks all vessel paths from a user-specified catheter position to one or more tumours, generating a vessel tree. The algorithm consists of two main stages. Firstly, the hepatic arterial tree is segmented and then a path is traversed through this segmentation mask, obtaining the final tracked vessel tree.

The vessel tree output can be used for the planning and guidance of a TACE procedure, providing a direct route from catheter to tumour for an intravenous chemotherapy injection to follow. Additionally, identifying all of the feeder vessels provides the locations to ensure complete embolisation, blocking blood entering or leaving the tumour to limit tumour growth and preventing the chemotherapeutic agent from being washed away.

OpenRib is an application designed to provide an automatically rendered, unfolded, unobstructed view of the entire ribcage to speed up fracture detection in trauma cases.

The application allows the user to see the unfolded ribs and standard clinical views for reference. There are two main steps: ribcage segmentation and ribcage unfolding. The unfolding technique we developed preserves the relative size and location of the ribs and surrounding tissue, providing a natural anatomical reference for the reader. This application unifies complex image analysis tech with advanced visualisation rendering methods, the latter developed by our colleagues in the Visualisation group.

OpenRib tech in use

What our people say

“I feel very lucky to be working with a group of such highly skilled individuals from both technical and clinical backgrounds. More so, the team has created a friendly environment, allowing open research discussion which lead to constructive collaborations on a range on interesting real-world world challenges!”

Catherine, Scientist

“In getting the opportunity to work with clinical collaborators and get direct custom feedback we get a better understanding of the challenges and pressures faced in making life changing decisions. Having the opportunity to investigate, develop and present novel Decision Support solutions that can make an impact on the confidence level on making such decisions is extremely rewarding.”

Alistair, Team Lead