update data dir

taigw · taigw · commit 8501e4aaeb1d · 2022-08-19T12:01:01.000+08:00
diff --git a/segmentation/JSRT/README.md b/segmentation/JSRT/README.md
@@ -6,14 +6,14 @@
 In this example, we use 2D U-Net to segment the lung from X-Ray images. First we edit the configuration file for training and testing. During training, we use tensorboard to observe the performance of the network at different iterations. We then apply the trained model to testing images and obtain quantitative evaluation results.
 
 ## Data and preprocessing
-1. Download the `PyMIC_data` and the images are saved in `PyMIC_data/JSRT`. The orginal dataset is available at the [JSRT website][jsrt_link]. It consists of 247 chest radiographs. We have preprocessed the images by resizing them to 256x256 and extracting the lung masks for the segmentation task. If interested, you can read `image_convert.py` for details of the preprocessing.  
+1. The orginal dataset is available at the [JSRT website][jsrt_link]. It consists of 247 chest radiographs. We have preprocessed the images by resizing them to 256x256 and extracting the lung masks for the segmentation task. The images are available at `PyMIC_data/JSRT`. If interested, you can read `image_convert.py` for details of the preprocessing.  
 
-5. Set `JSRT_root` according to your computer in `write_csv_files.py` and run `python write_csv_files.py` to randomly split the 247 images into training (180 images), validation (20 images) and testing (47 images) sets. The output csv files are saved in `config`.
+5. Run `python write_csv_files.py` to randomly split the 247 images into training (180 images), validation (20 images) and testing (47 images) sets. The output csv files are saved in `config`.
 
 [jsrt_link]:http://db.jsrt.or.jp/eng.php
 
 ## Training
-1. Edit `config/unet.cfg` by setting the value of `root_dir` as your `JSRT_root`. Then start to train by running:
+1. Start to train by running:
  
 ```bash
 pymic_run  train config/unet.cfg
diff --git a/segmentation/JSRT/config/evaluation.cfg b/segmentation/JSRT/config/evaluation.cfg
@@ -3,7 +3,7 @@ metric = dice
 label_list = [255]
 organ_name = lung
 
-ground_truth_folder_root  = /home/disk2t/projects/PyMIC_project/PyMIC_data/JSRT
+ground_truth_folder_root  = ../../PyMIC_data/JSRT
 segmentation_folder_root  = result
 evaluation_image_pair     = ./config/jsrt_test_gt_seg.csv
 
diff --git a/segmentation/JSRT/config/unet.cfg b/segmentation/JSRT/config/unet.cfg
@@ -3,7 +3,7 @@
 tensor_type = float
 
 task_type = seg
-root_dir  = /home/disk2t/projects/PyMIC_project/PyMIC_data/JSRT
+root_dir  = ../../PyMIC_data/JSRT
 train_csv = config/jsrt_train.csv
 valid_csv = config/jsrt_valid.csv
 test_csv  = config/jsrt_test.csv
diff --git a/segmentation/JSRT/write_csv_files.py b/segmentation/JSRT/write_csv_files.py
@@ -68,7 +68,7 @@ def get_evaluation_image_pairs(test_csv, gt_seg_csv):
 if __name__ == "__main__":
     # create cvs file for JSRT dataset
     # edit JSRT_root based on your system
-    JSRT_root   = '/home/disk2t/projects/PyMIC_project/PyMIC_data/JSRT'
+    JSRT_root   = '../../PyMIC_data/JSRT'
     output_file = 'config/jsrt_all.csv'
     fields      = ['image', 'label']
     create_csv_file(JSRT_root, output_file, fields)
diff --git a/segmentation/JSRT2/README.md b/segmentation/JSRT2/README.md
@@ -13,7 +13,7 @@ We also write a customized main function in `jsrt_net_run.py` so that we can com
 1. We use the same dataset as in the the `JSRT` example. 
 
 ## Training
-1. Edit `config/mynet.cfg` by setting the value of `root_dir` as your `JSRT_root`, and start to train by running:
+1. Start to train by running:
  
 ```bash
 python net_run_jsrt.py train config/mynet.cfg
diff --git a/segmentation/JSRT2/config/evaluation.cfg b/segmentation/JSRT2/config/evaluation.cfg
@@ -3,7 +3,7 @@ metric = dice
 label_list = [255]
 organ_name = lung
 
-ground_truth_folder_root  = /home/disk2t/projects/PyMIC_project/PyMIC_data/JSRT
+ground_truth_folder_root  = ../../PyMIC_data/JSRT
 segmentation_folder_root  = result
 evaluation_image_pair     = ../JSRT/config/jsrt_test_gt_seg.csv
 
diff --git a/segmentation/JSRT2/config/mynet.cfg b/segmentation/JSRT2/config/mynet.cfg
@@ -3,7 +3,7 @@
 tensor_type = float
 
 task_type = seg
-root_dir  = /home/disk2t/projects/PyMIC_project/PyMIC_data/JSRT
+root_dir  = ../../PyMIC_data/JSRT
 train_csv = ../JSRT/config/jsrt_train.csv
 valid_csv = ../JSRT/config/jsrt_valid.csv
 test_csv  = ../JSRT/config/jsrt_test.csv
diff --git a/segmentation/fetal_hc/README.md b/segmentation/fetal_hc/README.md
@@ -7,12 +7,12 @@ In this example, we use 2D U-Net to segment the fetal brain from ultrasound imag
 
 ## Data and preprocessing
 1. We use the [HC18][hc18_link] dataset for this example. The images are ready to use in `PyMIC_data/Fetal_HC`. There are 999 2D ultrasound images and their annotations in the training set.  The original annotations are contours, and we have converted them into binary masks for segmentation. (The conversion code is in `get_ground_truth.py`)
-2. Set `HC_root` according to your computer in `write_csv_files.py` and run `python write_csv_files.py` to randomly split the official HC18 training set into our own training (780 images), validation (70 images) and testing (149 images) sets. The output csv files are saved in `config`.
+2. Run `python write_csv_files.py` to randomly split the official HC18 training set into our own training (780 images), validation (70 images) and testing (149 images) sets. The output csv files are saved in `config`.
 
 [hc18_link]:https://hc18.grand-challenge.org/
 
 ## Training
-1. Edit `config/unet.cfg` by setting the value of `root_dir` as your `HC_root`. Then start to train by running:
+1. Start to train by running:
  
 ```bash
 pymic_run train config/unet.cfg
diff --git a/segmentation/fetal_hc/config/evaluation.cfg b/segmentation/fetal_hc/config/evaluation.cfg
@@ -3,6 +3,6 @@ metric = dice
 label_list = [255]
 organ_name = brain
 
-ground_truth_folder_root  = /home/disk2t/projects/PyMIC_project/PyMIC_data/Fetal_HC
+ground_truth_folder_root  = ../../PyMIC_data/Fetal_HC
 segmentation_folder_root  = result
 evaluation_image_pair     = ./config/fetal_hc_test_gt_seg.csv
diff --git a/segmentation/fetal_hc/config/unet.cfg b/segmentation/fetal_hc/config/unet.cfg
@@ -3,7 +3,7 @@
 tensor_type = float
 
 task_type = seg
-root_dir  = /home/disk2t/projects/PyMIC_project/PyMIC_data/Fetal_HC
+root_dir  = ../../PyMIC_data/Fetal_HC
 train_csv = config/fetal_hc_train.csv
 valid_csv = config/fetal_hc_valid.csv
 test_csv  = config/fetal_hc_test.csv
diff --git a/segmentation/fetal_hc/write_csv_files.py b/segmentation/fetal_hc/write_csv_files.py
@@ -67,7 +67,7 @@ def get_evaluation_image_pairs(test_csv, gt_seg_csv):
 
 if __name__ == "__main__":
     # create cvs file for training set
-    HC_root     = '/home/disk2t/projects/PyMIC_project/PyMIC_data/Fetal_HC'
+    HC_root     = '../../PyMIC_data/Fetal_HC'
     output_file = 'config/fetal_hc_all.csv'
     create_csv_file(HC_root, output_file)
 
diff --git a/segmentation/prostate/README.md b/segmentation/prostate/README.md
@@ -5,12 +5,12 @@ In this example, we use 3D U-Net with deep supervision to segment the prostate f
 
 ## Data 
 1. We use the [Promise12][promise12_link] dataset for this example. The preprocessed images are available in `PyMIC_data/Promise12`. We have resampled the original images into a uniform resolution and cropped them to a smaller size. The code for preprocessing is in  `preprocess.py`.
-2. Open `write_csv_files.py` and set `data_dir` accordingly, such as `PyMIC_data/Promise12/preprocess`. Run `python write_csv_files.py` to randomly split the dataset into our own training (35 images), validation (5 images) and testing (10 images) sets. The output csv files are saved in `config/data`.
+2. Run `python write_csv_files.py` to randomly split the dataset into our own training (35 images), validation (5 images) and testing (10 images) sets. The output csv files are saved in `config/data`.
 
 [promise12_link]:https://promise12.grand-challenge.org/
 
 ## Training
-1. Edit `config/unet3d.cfg` by setting the value of `root_dir` as your `PyMIC_data/Promise12/preprocess`. Then start to train by running:
+1. Start to train by running:
  
 ```bash
 pymic_run train config/unet3d.cfg
diff --git a/segmentation/prostate/config/evaluation.cfg b/segmentation/prostate/config/evaluation.cfg
@@ -3,7 +3,7 @@ metric = dice
 label_list = [1]
 organ_name = prostate
 
-ground_truth_folder_root  = /home/disk2t/projects/PyMIC_project/PyMIC_data/Promise12/preprocess
+ground_truth_folder_root  = ../../PyMIC_data/Promise12/preprocess
 segmentation_folder_root  = result
 evaluation_image_pair     = ./config/data/image_test_gt_seg.csv
 
diff --git a/segmentation/prostate/config/unet3d.cfg b/segmentation/prostate/config/unet3d.cfg
@@ -3,7 +3,7 @@
 tensor_type = float
 
 task_type = seg 
-root_dir  = /home/disk2t/projects/PyMIC_project/PyMIC_data/Promise12/preprocess
+root_dir  = ../../PyMIC_data/Promise12/preprocess
 train_csv = config/data/image_train.csv
 valid_csv = config/data/image_valid.csv
 test_csv  = config/data/image_test.csv
diff --git a/segmentation/prostate/write_csv_files.py b/segmentation/prostate/write_csv_files.py
@@ -79,7 +79,7 @@ def get_evaluation_image_pairs(test_csv, gt_seg_csv):
 if __name__ == "__main__":
     # create cvs file for promise 2012
     # set data_dir based on your own system
-    data_dir    = '/home/disk2t/projects/PyMIC_project/PyMIC_data/Promise12/preprocess'
+    data_dir    = '../../PyMIC_data/Promise12/preprocess'
     fields      = ['image', 'label']
     output_file = 'config/data/image_all.csv'
     create_csv_file(data_dir, output_file, fields)
