README.md

U-net-2D-EM-segmentation

using keras

aim to ISBI Challenge: Segmentation of neuronal structures in EM stacks,http://brainiac2.mit.edu/isbi_challenge/

you can download the data from the website or find data in the data folder

Pre-processing

The images are 3-D volume tiff, you should transfer the stacks into images first.of course the images in the data folder have already be transformed into 2D images. The data for training contains 30 512*512 images, which are far not enough to feed a deep learning neural network.
however without data augmentation we can still get a good result.
You should first run the augmentation.py to do augmentation, and then data.py to read data,then the unet.py to fit model
remember to change the file location in both files and create folders called results,aug,seg,listed below.

unet/
  data/
    train/
        0.tif
        1.tif
        ...
    label/
        0.tif
        1.tif
        ...
    test/
        0.tif
        1.tif
        ...
  results/
        #for augmentation and segmentation
        aug/
           train/
           0.tif
           1.tif
           ...
           label/
           0.tif
           1.tif
           ...
        image.npy
        seg/#results for the images in the test folder
           0.jpg
           1.jpg
           ...