Image Analysis

  • Image Analysis
  • Technology
  • The Image Analysis team is a long established team of scientific researchers, software engineers and clinical specialists who are all responsible for the delivery of component software on a variety of projects and are also experts in the support of component software already out in the field.

    The team works with international clinical collaborators on image analysis research and publications and with colleagues in Japan, USA and Europe to deliver on a variety of projects. These range from working on raw data from medical scanners to support in clinical decisions and new solutions in machine learning.

    To support them in their efforts the Image Analysis team have the opportunity to attend medical image analysis and machine learning courses and conferences, visit clinical sites and also have an internal training strategy to keep up to date with new methods, ideas and technology.

    Recent project releases have included component software in interventional oncology, coronary analysis and follow up registration.

    Here’s what our employees have to say:

    “We are involved in every aspect of component development and have the opportunity to transform new technology in it’s infancy to a sophisticated clinical solution to a real clinical problem.” Marco, Technology Team Lead

    “It’s just great working with smart, conscientious people on tech that can really help people.” Joseph, Scientist

    “I feel that the team is given a fair amount of autonomy to get on with developing applications as we see fit, including defining our own processes and trying things out like mob programming.” Dave, Senior Software Engineer

    “It’s great to have the freedom to explore new ideas and work on a variety of projects with other teams.” Kay, Software Team Lead

  • See below the latest innovations from the Image Analysis team

    Deep Learning for Ischemic Stroke

    This is a prototype application that uses a deep learning algorithm to detect signs of early ischemic change in non-contrast CT head images.  The algorithm is trained with various examples of patient data and at inference time highlights detections of; parenchymal hypoattenuation, loss of grey matter/white differentiation and sulcal effacement as well as the hyperdense artery sign which is associated with thromboembolic material in the vessel lumen.  Another algorithm estimates the ASPECTS territories and calculates the corresponding Alberta Stroke Program Early CT Score (ASPECTS) that helps to categorise the severity of the suspected stroke.  Due to the treatment of stroke being time critical we seek to improve clinical workflow and efficiency of stroke treatment.


    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.  We have also published a MICCAI paper on the underlying the algorithm.  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.
    The OpenRib application also demonstrated usefulness to identify other musculoskeletal conditions such as scoliosis, calcified cartilage, bone tumours.