https://github.com/ternaus/TernausNet

UNet model with VGG11 encoder pre-trained on Kaggle Carvana dataset
https://github.com/ternaus/TernausNet

image-segmentation pytorch

Last synced: about 1 month ago
JSON representation

UNet model with VGG11 encoder pre-trained on Kaggle Carvana dataset

• Host: GitHub
• URL: https://github.com/ternaus/TernausNet
• Owner: ternaus
• License: mit
• Created: 2017-12-30T03:13:50.000Z (almost 7 years ago)
• Default Branch: master
• Last Pushed: 2022-10-26T02:20:58.000Z (about 2 years ago)
• Last Synced: 2024-11-21T04:33:57.446Z (about 1 month ago)
• Topics: image-segmentation, pytorch
• Language: Python
• Homepage: https://arxiv.org/abs/1801.05746
• Size: 81.9 MB
• Stars: 1,048
• Watchers: 31
• Forks: 246
• Open Issues: 10
• Metadata Files:
  • Readme: README.md
  • Funding: .github/FUNDING.yml
  • License: LICENSE

Awesome Lists containing this project

README

        
# TernausNet: U-Net with VGG11 Encoder Pre-Trained on ImageNet for Image Segmentation

By [Vladimir Iglovikov](https://www.linkedin.com/in/iglovikov/) and [Alexey Shvets](https://www.linkedin.com/in/shvetsiya/)

# Introduction

TernausNet is a modification of the celebrated UNet architecture that is widely used for binary Image Segmentation. For more details, please refer to our [arXiv paper](https://arxiv.org/abs/1801.05746).

![UNet11](https://habrastorage.org/webt/hu/ji/ir/hujiirvpgpf7eswq88h_x7ahliw.png)

![loss_curve](https://habrastorage.org/webt/no/up/xq/noupxqqk_ivqwv3e7btyxtemt0m.png)

Pre-trained encoder speeds up convergence even on the datasets with a different semantic features. Above curve shows validation Jaccard Index (IOU) as a function of epochs for [Aerial Imagery](https://project.inria.fr/aerialimagelabeling/)

This architecture was a part of the [winning solutiuon](http://blog.kaggle.com/2017/12/22/carvana-image-masking-first-place-interview/) (1st out of 735 teams) in the [Carvana Image Masking Challenge](https://www.kaggle.com/c/carvana-image-masking-challenge).

# Installation

```bash

pip install ternausnet

```

# Citing TernausNet

Please cite TernausNet in your publications if it helps your research:

```

@ARTICLE{arXiv:1801.05746,

         author = {V. Iglovikov and A. Shvets},

          title = {TernausNet: U-Net with VGG11 Encoder Pre-Trained on ImageNet for Image Segmentation},

        journal = {ArXiv e-prints},

         eprint = {1801.05746},

           year = 2018

        }

```

# Example of the train and test pipeline

https://github.com/ternaus/robot-surgery-segmentation