zyl200846 / 3d-unet-pytorch-implementation

The implementation of 3D-UNet using PyTorch

3D-UNet-PyTorch-Implementation

This is the implementation of 3D UNet Proposed by Özgün Çiçek et al., for details please refer to: 3D U-Net: Learning Dense Volumetric Segmentation from Sparse Annotation.

Dataset used: Soft-tissue-Sarcoma, the dataset I used has been processed by other people and due to some reasons I cannot share it here. If you want to use it for you work, please refer to the MATLAB code provided and convert the images into proper format for further processing and analysis.

Description

Here describes the function of each file and folder

Training results

Here shows the training results including training loss and mIOU.

Note: Due to the limited hardward resources, the code are not yet successfuuly tested. WIP (BUSY recently, the code might be updated a litttle bit late. BUT, I am working on it constantly).

3d-unet-pytorch-implementation's Issues

how to run functions

so i've got a set of 3D images (128, 128, 128) (both images and their appropriate masks). I want to run this implementation of UNet on it. How would I do this? They're .nii files so i've coded up my on dataset loader for it. I'm a bit confused on where to get started as there aren't enough reference resources.

Where unet3d_model.tmp?

I ran "train.py" but the error was happended.
The error is
ModuleNotFoundError: No module named 'unet3d_model.tmp'

I can't find "tmp" in unet3d_model.
Where is "tmp"?

Sparse Annotation

您好，请问您的代码有实现稀疏标注数据的训练么？