sct-pipeline / contrast-agnostic-softseg-spinalcord

Contrast-agnostic spinal cord segmentation project with softseg

MIT License

4 stars 3 forks source link

Test contrast-agnostic model in Marseille 7T T2S data #74

Open Nilser3 opened 9 months ago

Nilser3 commented 9 months ago

Description

Issue to test the contrast-agnostic model in Marseille 7T T2*-w data

Data:

Shape: 128 x 128 and Z variable according to the number of slices
Resolution in plane: 0.18x0.18 All images are centered and cropped

Running inference

Based on @naga-karthik 's script

python monai/run_inference_single_image.py --path-img IMAGE --chkp-path  /mnt/duke/temp/muena/contrast-agnostic/final_monai_model/nnunet_nf\=32_DS\=1_opt\=adam_lr\=0.001_AdapW_CCrop_bs\=2_64x192x320_20230918-2253/ --path-out test_t2s --device cpu

Legend of the masks

SC-manual : GT
7t_multiclass_sc : sct_deepseg -task seg_gm_sc_7t_t2star
pred_bin : contrast-agnostic output binarized (thr 0.5001)

Categories

HC : Sub 01 to 35
MS : Sub 36 to 50
ALS : Sub 51 to 78