Test MS segmentation models in 3T 7T MP2RAGE from Marseille

[Nilser3]

Model to test: model_seg_ms_mp2rage

release: r20230210

Preprocessing based on script: https://github.com/ivadomed/model_seg_ms_mp2rage/issues/63#issue-1562548573

• [x] 3T MP2RAGE ( r = 0.9 x 0.9 x 0.9 )

Related in https://github.com/ivadomed/model_seg_ms_mp2rage/issues/63#issuecomment-1568903640

Example of automatic segmentation:

DICE Score values (N=40)

Continuous Dice Coefficient for "soft masks" based on this implementation and the Classic Dice between the GT and binarized masks (thresholding at 1/5 = 0.2) because we have 5 segmentation models.

Analysis of subjects with Dice = 0

7T MP2RAGE

• [x] isotropic images ( r = 0.7 x 0.7 x 0.7 )

• [x] isotropic images resampled to 1mm isotropic

• [x] anisotropic images (Same subject) ( r = 0.4 x 0.4 x 4 )

• [x] anisotropic images resampled to 1mm isotropic

subject with 3 lesions, 2 were segmented, one was not detected

[jcohenadad]

Thank you for the thorough report @Nilser3, I find it very encouraging!