About the project
Three dimensional images of the foot taken under loading conditions can provide a valuable clinical tool for the assessment of bone alignment related complaints. However, as these images have to be taken whilst a person is standing, specialised scanners are required to collect the image data. With limited availability of the required specialised equipment, most diagnostic decisions still have to be made based on traditional images, such as weightbearing two dimensional projective radiographic images or non-weightbearing three-dimensional X-ray computed tomography (CT) images, which can be generated with equipment readily available in most clinical settings.
This PhD project will explore the feasibility of combining information from several weightbearing two-dimensional projective X-ray images with non-weightbearing three-dimensional tomographic data to extract the clinically salient diagnostic information. Working closely with orthopaedic surgeons, the project is anticipated to use both simulated as well as real X-ray image data in order to develop advanced image processing and computer vision algorithms to combine information from the two modalities. Utilising the latest advances in machine learning, the project aims to overcome two fundamental challenges, 1) the identification of the unknown alignment of the two-dimensional projective x-ray images relative to the X-ray imaging system and 2) the identification of key anatomical landmarks in each of the images that will allow for the precise alignment of the different anatomical structures in each of the imaging conditions.
