Context

• Time consuming manual segmentation task of knee bones.
• Difficult segmentation of bones affected by typical pathologies.
• Need for human intervention all along the segmentation process.
• Difficult choice of the prosthesis shape before the surgery.

Solution

A semi automated tool for segmentation based on a statistical approach and a computer assisted segmentation tool based on geodesic radial basis functions:

• Fully automated solution for the standard cases (University of Basel).
• Manual segmentation refinement for patients with pathologies affecting the automatic process (HEIG-VD).
• 3D segmentation process to replace the current slice-by-slice manual approach.

Advantages

• More precise segmentation of the articular bones.
• Reduced segmentation time (man-hours).
• Simplify the manual segmentation of complex cases.
• Improve the final segmentation.