Introduction to research - Analysis by Supervised Learning of CT images of pelvic fractures
What is that ?
I had the chance to work on a real project during my studies. It was about the analysis of CT images of pelvic fractures.
I worked one afternoon per week during one semester in the TIMC laboratory in the GMCAO team.
Abstract
Acetabular fracture (cavity in the pelvis where the head of the femur is located), generally require a CT scan to locate and classify the fractures according to their type, according to the Letournel classification. It is then possible to segment the bone fragments in order to visualize them in 3D, but above all to perform a surgical reduction simulation in order to validate or not the proposed surgical planning. Currently, such a segmentation is semi-automatic and takes between 2 and 3 hours, which is too long for a routine use of the simulation tool. The objective is then to exploit the new methods, and in particular the novelties in order to automate this segmentation process for which a good precision is required given its medical application.
The paper I wrote about my work (in French) (unpublished)
Publication
This work has been continued and published in the SPIE Medical Imaging journal
Towards a learning-based CT segmentation of acetabular fractures