/Nested-VNet3D

Nested VNet3D Model for KiTS19——2019 Kidney Tumor Segmentation Challenge

Primary LanguagePython

Nested VNet3D Model for KiTS19——2019 Kidney Tumor Segmentation Challenge

This is an example of Nested VNet3D.(re)implemented the model with tensorflow framework in the paper of "Zongwei Zhou, Md Mahfuzur Rahman Siddiquee, Nima Tajbakhsh, Jianming Liang.(2018) UNet++: A Nested U-Net Architecture for Medical Image Segmentation"

Prerequisities

The following dependencies are needed:

  • numpy >= 1.11.1
  • opencv-python >=3.3.0
  • tensorflow-gpu ==1.8.0
  • pandas >=0.20.1
  • scikit-learn >= 0.17.1

How to Use

Preprocess

follow KiTS19——2019 Kidney Tumor Segmentation Challenge project

  • the VNet3D++ model

  • the skip pathway function

Prerequisities

UNet++: A Nested U-Net model can find from here:https://github.com/MrGiovanni/UNetPlusPlus

Result

1、Kidney Segmentation

  • the train loss

  • 200-209case dice value and result

  • segmentation result can find in the video of VNet++.avi

Contact