Update README.md

README.md

@@ -5,6 +5,30 @@ This repository provides the nnUNet v2 checkpoint for multi-organ lesion segment
 
 The model was trained on a large-scale multi-center CT dataset with expert-annotated lesion masks across multiple organs (liver, pancreas, kidney, colon).
 
+## Label Definitions
+```bash
+"labels": {
+        "background": 0,
+        "liver_segment_1": 1,
+        "liver_segment_2": 2,
+        "liver_segment_3": 3,
+        "liver_segment_4": 4,
+        "liver_segment_5": 5,
+        "liver_segment_6": 6,
+        "liver_segment_7": 7,
+        "liver_segment_8": 8,
+        "pancreas_head": 9,
+        "pancreas_body": 10,
+        "pancreas_tail": 11,
+        "kidney_left": 12,
+        "kidney_right": 13,
+        "colon": 14,
+        "liver_lesion": 15,
+        "pancreatic_lesion": 16,
+        "kidney_lesion": 17,
+        "colon_lesion": 18
+    }
+```
 
 
 ## 1. Installation
@@ -17,53 +41,40 @@ pip install -e .
 
 ## 2. Inference 
 
-#### 2.1 Convert dataset (CT scans) to nnUNet naming format and 
 
+#### 2.1 Create a folder (e.g., nnUNet_eval) and rename all .ct files and put them in it.
 
+```bash
+nnUNet_eval/
+ └── Dataset1351/
+      └── imagesTs/
+           ├── ct001_0000.nii.gz
+           ├── ct002_0000.nii.gz
+```
 
+#### 2.2 Place the checkpoint into
 
 ```bash
-$OUTPUT_BDMAP_PATH
-│── BDMAP_0000001_0000.nii.gz
-│── BDMAP_0000002_0000.nii.gz
-└── ...
+nnUNet_results/
+└── Dataset1351/nnUNetTrainer__nnUNetResEncUNetLPlans__3d_fullres/fold_all/
 ```
 
-**Download the best checkpoint (refer to shell script in PanTS)**
+#### 2.3 Run inference
 
 ```bash
-bash download_checkpoints.sh
+export nnUNet_N_proc_DA=36
+export nnUNet_results="./nnUNet_results"
+export nnUNet_predictions="./nnUNet_predictions"
+export nnUNet_eval="./nnUNet_eval"
 
-wget http://www.cs.jhu.edu/~zongwei/model/xzhou120_DAP_Atlas_Full.tar.gz
-tar -xzvf xzhou120_DAP_Atlas_Full.tar.gz
+GPU_ID=0 
+DATASET=1351
 
-wget http://www.cs.jhu.edu/~zongwei/model/xzhou120_DAP_Atlas_Mini.tar.gz
-tar -xzvf xzhou120_DAP_Atlas_Mini.tar.gz
+TRAINER=nnUNetTrainer 
 
-wget http://www.cs.jhu.edu/~zongwei/model/xzhou120_Lymph_node.tar.gz
-tar -xzvf xzhou120_Lymph_node.tar.gz
-```
+CUDA_VISIBLE_DEVICES=$GPU_ID nnUNetv2_predict -d $DATASET -i $nnUNet_eval/ -o $nnUNet_predictions/ -tr $TRAINER -d $DATASET -c 3d_fullres -f all -p nnUNetResEncUNetLPlans --continue_prediction 
 
-##### 2.2 Run the inference process (Single GPU with GPUID=0)
-```bash
-GPUID=1
-ID_TO_PROCESS=batch1.csv
-CHECKPOINT_TO_RUN=checkpoint1.csv
-bash run.sh $GPUID $ID_TO_PROCESS $CHECKPOINT_TO_RUN
 ```
 
-When completed, the results will be save as
 
-```bash
-$INTERMEDIATE_OUTPUT
-BDMAP_00000001/
-├── predictions/
-│   ├── V-Rad_0001/
-│   │   ├── class_0001.nii.gz
-│   │   ├── class_0003.nii.gz
-│   │   └── class_0004.nii.gz
-│   └── V-Rad_0002/
-│   │   ├── class_0002.nii.gz
-│   │   ├── class_0004.nii.gz
-│   │   └── class_0005.nii.gz
-```
+