Home

DeepBacs

Analyzing bacterial bioimages with open-source deep learning approaches

Deep learning is a powerful tool in many areas. In the last years, it also gained large popularity in the field of bioimage analysis. Yet, bacteriologists make only limited use of this technology, although there is great potential. To leverage the use of DL in microbiology, we created several datasets of bacterial bioimages and tested DL networks for different image analysis tasks.

Our work is based on the ZeroCostDL4Mic platform, as it provides simple access to different DL networks and gives you free GPU computing power! For this it employs the Google Colab platform.

Here's an overview on the tasks that we perform using open-source DL approaches:

Scale bars are 2 µm

We are always happy for new ideas and examples. Please feel free to contact us if you want to contribute data and models or want to discuss any useful applications of DL for microbiology!

---

List of tasks with links to the respective pages:

• Segmentation
• Object detection
• Denoising
• Artificial labeling
• Prediction of super-resolution images

Notebooks used in our work:

All notebooks are implementations provided by the ZeroCostDL4Mic platform. Further documentation can be found on the ZeroCostDL4Mic wiki)

Network	Paper(s)	Task	Link to DeepBac example training and test dataset	Direct link to notebook in Colab
U-Net (2D)	here and here	Segmentation	add segmentation link here
StarDist (2D)	here and here	Nuclei segmentation	add segmentation link here
SplineDist	here	Instance segmentation	add segmentation link here
Noise2Void (2D)	here	Denoising	[B. subtilis FtsZ](add link) and [E. coli nucleoid](add link)
CARE (2D)	here	Denoising	[E. coli nucleoid](add link)
Label-free prediction (fnet) 2D	here	Artificial labelling	[Artificial labelling](add link)
pix2pix	here	Paired Image-to-Image Translation	[B. subtilis segmentation](add link)
YOLOv2	here	Object detection (bounding boxes)	[Growth stage](add link) and [Antibiotic phenotyping](add link)

Tools

Network	Paper(s)	Task	Link to example training and test dataset	Direct link to the notebook in Colab
Augmentor	here	Image augmentation	None
Quality Control	Available soon	Error mapping and quality metrics estimation	None

Contributors

• Christoph Spahn
• Romain F. Laine
• Pedro Matos Pereira
• Lucas von Chamier
• Seamus Holden
• Mia Conduit
• Mariana de Gomes Pinho
• Guillaume Jacquemet
• Mike Heilemann
• Ricardo Henriques

You can find our preprint here ## add preprint link here