Changed file structure. Updated to Python3/Keras2

README.md

@@ -48,27 +48,30 @@ This tutorial depends on the following libraries:
 
 Also, this code should be compatible with Python versions 2.7-3.5.
 
+You can also set up a virtual enviroment and run your code from there:
+* Create virtual enviroment folder: ```virtualenv -p python3 venv```
+* Activate your virtual enviroment: ```source vev/bin/activate```
+* Install packages needed in enviroment: ```pip3 install -r requirements.txt```
+* Check installed packages: ```pip3 freeze```
+
 ### Prepare the data
 
 First transfer 3D volume tiff to 30 512*512 images.
 
 To feed the unet, data augmentation is necessary.
 
-An [image deformation](http://faculty.cs.tamu.edu/schaefer/research/mls.pdf) method is used, the code is 
+An [image deformation](http://faculty.cs.tamu.edu/schaefer/research/mls.pdf) method is used, the code is
 
 availabel in this [repository](https://github.com/cxcxcxcx/imgwarp-opencv).
 
 
-
-
 ### Define the model
 
 * Check out ```get_unet()``` in ```unet.py``` to modify the model, optimizer and loss function.
 
 ### Train the model and generate masks for test images
 
-* Run ```python unet.py``` to train the model.
-
+* Run ```python3 unet.py``` to train the model.
 
 After this script finishes, in ```imgs_mask_test.npy``` masks for corresponding images in ```imgs_test.npy```
 should be generated. I suggest you examine these masks for getting further insight of your model's performance.
@@ -81,6 +84,25 @@ Use the trained model to do segmentation on test images, the result is statisfac
 
 ![img/0label.png](img/0label.png)
 
+## Folder structure
+
+* ```./data```: Contains sample data which can be used to test unet.
+* ```./deform```: Contains 3 folders which are used when running your program. Images in ```./deform/train``` and ```./deform/label``` are used to train your network. After building your model images from ```./deform/test``` are used as inputs to produce results.
+* ```./npydata```: Contains the ```.npy``` files generated from ```.tif``` images from our deform folder.
+* ```./results```: Outputs from our unet are saved in this folder.
+
+## From clone to your first results
+
+* After cloning this project the first step is loading up the deform folder with the data you want to use to build and test your model. There is a sample dataset in the ```./data``` folder.
+
+* ```python3 data.py``` to generate ```.npy``` files from your input data. This should produce 3 files in your ```./npydata``` folder.
+
+* ```python3 unet.py``` to build your model. This should create a ```unet.hdf5``` file which is your model and can be loaded in your future programs. It also generates the results from the test dataset.
+
+* Additionally you can use ```python3 test_predict.py``` to generate results for different test sets.
+
+
+
 
 ## About Keras
 

data.py

@@ -1,12 +1,12 @@
 from keras.preprocessing.image import ImageDataGenerator, array_to_img, img_to_array, load_img
-import numpy as np 
+import numpy as np
 import os
 import glob
 #import cv2
 #from libtiff import TIFF
 
 class myAugmentation(object):
-	
+
 	"""
 	A class used to augmentate image
 	Firstly, read train image and label seperately, and then merge them together for the next process
@@ -15,7 +15,7 @@ class myAugmentation(object):
 	"""
 
 	def __init__(self, train_path="train", label_path="label", merge_path="merge", aug_merge_path="aug_merge", aug_train_path="aug_train", aug_label_path="aug_label", img_type="tif"):
-		
+
 		"""
 		Using glob to get all .img_type form path
 		"""
@@ -52,7 +52,7 @@ def Augmentation(self):
 		imgtype = self.img_type
 		path_aug_merge = self.aug_merge_path
 		if len(trains) != len(labels) or len(trains) == 0 or len(trains) == 0:
-			print "trains can't match labels"
+			print("trains can't match labels")
 			return 0
 		for i in range(len(trains)):
 			img_t = load_img(path_train+"/"+str(i)+"."+imgtype)
@@ -135,10 +135,10 @@ def splitTransform(self):
 
 class dataProcess(object):
 
-	def __init__(self, out_rows, out_cols, data_path = "../deform/train", label_path = "../deform/label", test_path = "../test", npy_path = "../npydata", img_type = "tif"):
+	def __init__(self, out_rows, out_cols, data_path = "./deform/train", label_path = "./deform/label", test_path = "./deform/test", npy_path = "./npydata", img_type = "tif"):
 
 		"""
-		
+
 		"""
 
 		self.out_rows = out_rows
@@ -208,7 +208,7 @@ def load_train_data(self):
 		imgs_mask_train = imgs_mask_train.astype('float32')
 		imgs_train /= 255
 		#mean = imgs_train.mean(axis = 0)
-		#imgs_train -= mean	
+		#imgs_train -= mean
 		imgs_mask_train /= 255
 		imgs_mask_train[imgs_mask_train > 0.5] = 1
 		imgs_mask_train[imgs_mask_train <= 0.5] = 0
@@ -222,7 +222,7 @@ def load_test_data(self):
 		imgs_test = imgs_test.astype('float32')
 		imgs_test /= 255
 		#mean = imgs_test.mean(axis = 0)
-		#imgs_test -= mean	
+		#imgs_test -= mean
 		return imgs_test
 
 if __name__ == "__main__":

requirements.txt

@@ -0,0 +1,21 @@
+bleach==1.5.0
+cycler==0.10.0
+enum34==1.1.6
+h5py==2.7.1
+html5lib==0.9999999
+Keras==2.1.2
+Markdown==2.6.10
+matplotlib==2.1.0
+numpy==1.13.3
+olefile==0.44
+Pillow==4.3.0
+protobuf==3.5.0.post1
+pyparsing==2.2.0
+python-dateutil==2.6.1
+pytz==2017.3
+PyYAML==3.12
+scipy==1.0.0
+six==1.11.0
+tensorflow==1.4.1
+tensorflow-tensorboard==0.4.0rc3
+Werkzeug==0.13

test_predict.py

@@ -13,4 +13,4 @@
 
 imgs_mask_test = model.predict(imgs_test, verbose=1)
 
-np.save('imgs_mask_test.npy', imgs_mask_test)
+np.save('./results/imgs_mask_test.npy', imgs_mask_test)