In this work, we study the utility of SinGAN—an unconditional generative model trained on a single image—for synthetic data generationacross different imaging tasks, namely, multi-modal MRI (Magnetic ResonanceImaging) data generation, brain tumour data generation, and histopathology imagesegmentation. These applications, built on SinGAN, could become a significantremedy to the data-deprived medical imaging tasks
The project has been forked from SinGAN
Project | Arxiv | CVF | Supplementary materials | Talk (ICCV`19)
Official pytorch implementation of the paper: "SinGAN: Learning a Generative Model from a Single Natural Image"
With SinGAN, you can train a generative model from a single natural image, and then generate random samples from the given image, for example:
SinGAN can be also used for a line of image manipulation tasks, for example:
This is done by injecting an image to the already trained model. See section 4 in the paper for more details.
If you use this code for your research, please cite our paper:
@inproceedings{rottshaham2019singan,
title={SinGAN: Learning a Generative Model from a Single Natural Image},
author={Rott Shaham, Tamar and Dekel, Tali and Michaeli, Tomer},
booktitle={Computer Vision (ICCV), IEEE International Conference on},
year={2019}
}
Our wonderful people 💻 :
- Amrit Kumar Jethi
- Rajkumar Vaghashiya
- Екатерина Неповинных
- Anagha Zachariah
- Madhu mithra K K
- Phuc Nguyen
The authors would like to thank the Eastern European Machine Learning Summer School(EEML’21) team, especially Viorica Patraucean, Razvan Pascanu, and Ferenc Huszar for the in-valuable knowledge and the opportunity to work together on this project.
The authors would like to thank the advisers Margarete Kattau, The Institute of Cancer Research,London, and Fedor Zolotarev, LUT University, Finland, for their guidance and support.