README.md

---
license: bsd-3-clause
tags:
- Confocal Fluorescence Microscopy
- Image Super-resolution
- Deep Learning
- Benchmark
---

# [SR-CACO-2: A Dataset for Confocal Fluorescence Microscopy Image Super-Resolution (NeurIPS2024)](https://arxiv.org/pdf/2406.09168.pdf)


by **Soufiane Belharbi<sup>1</sup>, Mara KM Whitford<sup>2,3</sup>, 
Phuong Hoang<sup>2</sup>, Shakeeb Murtaza<sup>1</sup>, Luke McCaffrey<sup>2,3,4</sup>, Eric Granger<sup>1</sup>**


<sup>1</sup>  LIVIA, Dept. of Systems Engineering, ETS Montreal, Canada
<br/>
<sup>2</sup>  Goodman Cancer Institute, McGill University, Montreal, Canada
<br/>
<sup>3</sup>  Dept. of Biochemistry, McGill University, Montreal, Canada
<br/>
<sup>4</sup>  Gerald Bronfman Dept. of Oncology, McGill University, Montreal,
Canada

<p align="center"><img src="patch-demo.png" alt="outline" width="60%"></p>

[![arXiv](https://img.shields.io/badge/arXiv-2406.09168-b31b1b.svg)](https://arxiv.org/pdf/2406.09168)
[![Github](https://img.shields.io/badge/github-sr--caco--2-brightgreen.svg)](https://github.com/sbelharbi/sr-caco-2)


## Abstract
Confocal fluorescence microscopy is one of the most accessible and widely used 
imaging techniques for the study of biological processes at the cellular and 
subcellular levels. Scanning confocal microscopy allows the capture of 
high-quality images from thick three-dimensional (3D) samples, yet suffers from 
well-known limitations such as photobleaching and phototoxicity of specimens 
caused by intense light exposure, which limits its use in some applications, 
especially for living cells. Cellular damage can be alleviated by changing 
imaging parameters to reduce light exposure, often at the expense of image 
quality. Machine/deep learning methods for single-image super-resolution 
(SISR) can be applied to restore image quality by upscaling lower-resolution 
(LR) images to produce high-resolution images (HR). These SISR methods have 
been successfully applied to photo-realistic images due partly to the abundance 
of publicly available datasets. In contrast, the lack of publicly available 
data partly limits their application and success in scanning confocal 
microscopy. In this paper, we introduce a large scanning confocal microscopy 
dataset named SR-CACO-2 that is comprised of low- and high-resolution image pairs 
marked for three different fluorescent markers. It allows to evaluate the 
performance of SISR methods on three different upscaling levels 
(X2, x34, x8). SR-CACO-2 contains the human epithelial cell line Caco-2 
(ATCC HTB-37), and it is composed of 2,200 unique images, captured with four 
resolutions and three markers, that have been translated in the form of 9,937 
patches for experiments with SISR methods. Given the new SR-CACO-2 dataset, 
we also provide benchmarking results for 16 state-of-the-art methods that are 
representative of the main SISR families. Results show that these methods have 
limited success in producing high-resolution textures, indicating that SR-CACO-2 
represents a challenging problem. The dataset is released under a Creative 
Commons license (CC BY-NC-SA 4.0), and it can be accessed freely. Our dataset, 
code and pretrained weights for SISR methods are publicly available: https://github.com/sbelharbi/sr-caco-2.

**Code: Pytorch 2.0.0**

## Citation:
```
@inproceedings{belharbi24-sr-caco-2,
  title={SR-CACO-2: A Dataset for Confocal Fluorescence Microscopy Image Super-Resolution},
  author={Belharbi, S. and Whitford, M.K.M. and Hoang, P. and Murtaza, S. and McCaffrey, L. and Granger, E.},
  booktitle={NeurIPS},
  year={2024}
}
```

<p align="center"><img src="nutrition-label.png" alt="nutrition label for SR-CACO-2 dataset" width="60%"></p>



## <a name="weights"> Pretrained weights (evaluation) </a>:
We provide the weights for all the models (135 models: 15 methods x 3 cells 
x 3 scales). Weights can be found at [Hugging Face](https://huggingface.co/sbelharbi/sr-caco-2) in the file [shared-trained-models.tar.gz](https://huggingface.co/sbelharbi/sr-caco-2/resolve/main/shared-trained-models.tar.gz?download=true).

The file [share-visualization-30-samples-test.zip](https://huggingface.co/sbelharbi/sr-caco-2/resolve/main/share-visualization-30-samples-test.zip?download=true) contains visual predictions on the test set.

The provided weights can be used to reproduce the reported results in the 
paper in the paper:
<p align="center"><img src="roi-perf.png" alt="roi performance" width="80%"></p>
<p align="center"><img src="full-img-perf.png" alt="full image performance" width="80%"></p>