msk-STAPLE icon indicating copy to clipboard operation
msk-STAPLE copied to clipboard

Published 20 hours ago • verified publisher icon modenaxe

[GIBOC] Small sesamoids bone of the foot

Open renaultJB opened this issue 5 years ago • 5 comments
trafficstars

This is more a question than an issue but would it be interesting to get rid of the small sesamoids bones of the foot. They could be removed from the foot triangulation using the new function TriSplitObjects.m that separated/split the unconnected triangulations from a general one. And sort them in a structure in descending order of volume. @modenaxe, would that be useful ?

FootSesam

renaultJB avatar Mar 19 '20 21:03 renaultJB

This is a nice feature. Maybe we could use it for removing fibula from tibia as well?

modenaxe avatar Apr 20 '20 14:04 modenaxe

I used this function to eliminate the fibula from the segmentations of methods that were not supposed to have the fibula in it (Miranda and GIBOK). @renaultJB if you confirm that GIBOK works better without Fibula (see this commit I will close the issue. I am asking because results made sense also with Fibula in the TIbia triangulations.

modenaxe avatar Apr 30 '20 08:04 modenaxe

The commit seems good. Giboc was designed to work without the fibula. The fibula will interfere in the identification of the most distal and medial part of the tibia (tip of the medial malleolus). On the proximal epiphysis it could interfere with the identification of the articular surface because the algorithm assumes the cross-section curves are unique. On situation to consider is the case were the tibia and fibula are not separated during segmentation and the triangulation "contains" both. This was the rationale of Kai 2014 not to remove the fibula from the tibia. May be a function to detect that the tibia triangulation is actually the [Tibia+Fibula] and force the usage of Kai method for the tibia could be interesting.

renaultJB avatar May 02 '20 10:05 renaultJB

Very good point. We can slice the tibia and if we find >= 2 areas but just one triangulation object then we force to use Kai. My approach was more relying on the user segmenting individual bones but I guess there is no harm in enforcing the correct method for this situation. I will deal with it and close the issue

modenaxe avatar May 04 '20 08:05 modenaxe

I haven't used it at the foot yet, this could probably be added in GIBOC/Calcn for the ankle paper. @renaultJB please remind me about that.

modenaxe avatar May 13 '20 09:05 modenaxe