Skip to content
Segmentation of medical images have brought a considerable impact on the diagnosis, medicine, and treatment. This segmentation of images helps the doctors in exploring the internal anatomy. There are many existing techniques based on cross-section images and X-Ray like Computed Tomography (CT), or Magnetic Resonance Imaging, or others like Singl…
Jupyter Notebook
Branch: master
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
README.md Update README.md Jul 18, 2019
Segthor_up.ipynb
giphy.gif Add files via upload Jul 18, 2019
giphy_test.gif Add files via upload Jul 18, 2019

README.md

Segthor19-using-ResU-net

The recent advances in the field of computer vision has led to the wide use of Convolutional Neural Networks (CNNs) in organ segmentation of computed tomography (CT) images. Image guided radiation therapy requires the accurate segmentation of organs at risk (OARs). In this paper, we propose a 2D U-Net network to automatically segment thoracic organs at risk in computed tomography (CT) images. The architecture consists of a down sampling path to capture features and a symmetric up sampling path to obtain precise localization. SegTHOR19 is a competition timed to the conference IEEE ISBI 2019 that addresses the problem of organs at risk segmentation in Computed Tomography (CT) images. In the SegTHOR19 challenge, 40 CT scans with 4 thoracic organs (i.e., esophagus, heart, trachea and aorta) were used for training [1]. We experimented with both 2D U-net and 2D U-Net with Resnet18 architecture to train the networks. Our best results were obtained by using 2D Convolutional U-Net with ResNet18.

For full demo of this project

YouTube video link -- https://youtu.be/jOHFZobAvwc

author: Mohit Asudani (IIIT MANIPUR)

Institution: Bennett University

email: m.asudani@iiitmanipur.ac.in

Tookit need Python 3, pytorch >= 1.0.1, fastai version 1.x.

The data path is organized like:
./dataset/train/Patient_01/GT.nii
./dataset/train/Patient_01/Patient_01.nii

GPU used : NVIDIA® DGX-1™ (8 gpu) COLAB GPU

DATA VISUALIZATION:

Dataset for training

Dataset for training

Data obtained by testing

Data obtained by testing



Please don't hesitate to contact me if you have any question about the data, method, or code !

You can’t perform that action at this time.