An Automated Segmentation Algorithm To identify Bones of The Hand and Wrist for Improved Structural Damage Detection
Keywords
Machine Learning, Computer Vision, Psoriatic Arthritis, Object Detection, Probabilistic
Psoriatic arthritis is a condition linked to psoriasis that can damage bones and joints, often in the hands and wrists. Doctors use X-rays to assess this damage, but scoring the images is time-consuming and requires specialist training. This study investigated a first step towards using artificial intelligence to help with this process. The model automatically found and outlined bones in hand and wrist X-rays, while also highlighting areas where it was less certain. The model was trained using expert-labelled images and performed close to human level, showing promise as an early step towards faster, more consistent and easier-to-interpret assessment of damage.


