Skip to content

Latest commit

 

History

History
37 lines (23 loc) · 1.18 KB

README.md

File metadata and controls

37 lines (23 loc) · 1.18 KB

DeepSlides

This study is based on Tensorflow implementation of pix2pix.

if you want to download the image dataset, please click here

Tech

DeepSlides uses a number of open source projects to work properly:

  • [Python]
  • [Tensorflow]
  • [Numpy]

How to use:

  • Download and unzip trained network for ki67 stained breast cancer. (to "code\artificialKi67")

  • Please create a png file. (\Testdata\xxx.png)

  • The size of the square image should be 256*n X 256*n where n=1,2,3,...

  • Draw the possitive and negative cells with green and red colors respectively.

  • Also you can use yellow color to represent possitive cells with staining problem.

  • Run createTissue.py --scale_size image size (256*n)

  • note: Please be sure that images sizes are same in that folder.

  • note: Please be sure that the annotated cells' area are realistic for 40x magnification

example image:

input mask

output image:

input mask

Todos

  • Web based client GUI