OsteophyteNets – Artificial Intelligence for detection bone formation in spine
Over one third of the UK population live with a musculoskeletal condition affecting the joints or spine. Amongst them, Psoriatic arthritis (PsA) and axial Spondyloarthritis (axSpA) are progressive and destructive forms of arthritis that affect 320,000 people in the UK. In the current diagnostic pathway, radiographs (X-ray imaging) remain the mainstay of damage assessment as they are accessible, quick, require minimal radiation and are low cost compared to other imaging modalities such Magnetic Resonance Imaging. Advances in Artificial Intelligence (AI) now open the opportunity to automate quantification of damage on radiographs. Machine Learning, and in particular Deep Learning (DL), has enabled artifact detection, grading medical imaging, and detection of pathological conditions.
The project will aim to investigate the state-of-the-art computer vision object detection methods that could be used for detection of new bone formation (osteophytes) in spine X-ray imaging.
The project may also involve modifying the existing object detection methods to implement a bespoke AI/DL model for osteophytes detection, and writing up the report with a future publication in mind.
8-12 weeks, from 19 June onward.
Primary supervisor available during the project