Extraction of an image mask from polygon data

Steven Boyd August 22, 2024 October 30, 2024

Project Description

Developing a new method with universal application for defining bone compartments…

Cortical bone mapping with global optimization produces incredible results for all types of CT data, including clinical CT and HR-pQCT. The result of the algorithm is a polygon dataset of the periosteal surface of the bone and scalars representing the local thickness (i.e., periosteal to endosteal) calculated by the algorithm. The calculation of cortical thickness can be summarized from the individual data points (i.e., means, standard deviations, localized measurements, etc).

https://github.com/Bonelab/Bonelab/blob/master/bonelab/cli/treece_thickness.py

However, the algorithm cannot produce an image mask. So the aim of this project is to develop additional functionality that outputs a binarized image with voxels labelled as cortical bone (127), trabecular bone (100), and background (0). The image will be the identical size to the input that is input. Likely it would be based on VTK stencil objects (similar to blRapidPrototype):

https://github.com/Bonelab/Bonelab/blob/master/bonelab/cli/RapidPrototype.py

When completed it will be validated on HR-pQCT of radius, tibia and knee, as well as QCT of abdomen and knee.

Scope

PhD/MSc side project
Summer student

Data Source

ARTININ
CaBHS

Resources

Treece surface model in Bonelab github