Aydin is a user-friendly, feature-rich, and fast image denoising tool that provides a number of self-supervised, auto-tuned, and unsupervised image denoising algorithms. Aydin handles from the get-go n-dimensional array-structured images with an arbitrary number of batch dimensions, channel dimensions, and typically up to 4 spatio-temporal dimensions.
It comes with Aydin Studio a graphical user interface to easily experiment with all the different algorithms and parameters available, a command line interface to run large jobs on the terminal possibly on powerfull remote machines, and an API for custom coding and integration into your scripts and applications.
And, of course, a simplified napari plugin is in the works! Notebooks for running on Collab are also planned.
Supported algorithms:
Currently Aydin supports two main families of denoisers: the first family consists of 'classical' denoising algorithms that leverage among other: frequency domain filtering, smoothness priors, low-rank representations, self-similarity. The second family consists of algorithms that leverage machine learning approaches such as convolutional neural networks (CNN) or gradient boosting (GB). In the Noise2Self paper we show that it is possible to calibrate any parameterised denoising algorithm, from the few parameters of a classical algorithm to the millions of weights of a deep neural network. We leverage and extend these ideas in Aydin to provide a variety of auto-tuned and trained high-quality image denoisers. Here is the list of currently available methods:
-
Low-pass filtering based algorithms:
- Butterworth denoiser (butterworth).
- Gaussian blur denoiser (gaussian).
- Gaussian-Median mixed denoiser (gm).
-
Optimisation-based with smoothness priors:
- Total-Variation denoising (tv)
- Harmonic prior (harmonic)
-
Low-rank representations:
- Denoising via sparse decomposition (e.g. OMP) over a fixed dictionary (DCT, DST, ...)
- Denoising via sparse decomposition (e.g. OMP) over a learned dictionary (Kmeans, PCA, ICA, SDL, ...)
-
Patch similarity:
- Non-Local Means denoising (nlm)
- BMnD (not available just yet but partly implemented!)
-
Machine learning based:
- Noise2Self-FGR: Noise2Self denoising via Feature Generation and Regression (FGR). We use specially crafted integral features. We have several variants that leverage different regressors: CatBoost(cb), lightGBM, linear, perceptron, random-forrest, support vector regression)
- Noise2Self-CNN: Noise2Self denoising via Convolutional Neural Networks (CNN). This is the original approach of Noise2Self. In our experience this is typically slower to train, and more prone to hallucination and residual noise than FGR.
-
Other:
- Lipschitz continuity denoising
Some methods actually do combine multiple ideas and so the classification above is not crisp. We recomend to try first a good baseline denoiser such as the Butterworth denoiser. If unsatisfied with the result and you have a powerfull computer with a recent NVIDIA graphics card then try the Noise2Self-FGR-cb denoiser. For detailed use-cases check here.
We have a long todo list of existing, modified, as well as original algorithms that we plan to add to Aydin in the next weeks and months. We will do so progressively as time allows. Stay tuned!
Documentation
Aydin documentation can be found here.
Installation of Aydin Studio
We recommend that users that are not familiar with python start with our user-friendly UI. Just download the version for your operating system:
The latest releases and standalone executables can be found here and detailed installation instructions of Aydin Studio for all three oerating systems can be found here
Installation of Aydin in Conda:
First download and install miniconda or anaconda.
Then create a conda environment to host aydin:
conda create -n aydin_env python=3.9
Activate the environment:
conda activate aydin_env
Install Aydin in the environment:
pip install aydin
CUDA dependencies:
If you have a NVIDIA CUDA Graphics card, you should install the CUDA toolkit:
conda install cudatoolkit
Mac specific dependencies:
For Macs (OSX) you first need to do:
brew install libomp
You can install Brew by following the instructions here.
How to run ?
Assuming that you have installed Aydin in an environment, you can:
Start Aydin Studio from the command line with:
aydin
Run the Command Line Interface (CLI) for denoising:
aydin denoise path/to/noisyimage
Get help on command line usage:
aydin -h
Recommended Environment
Linux
- Ubuntu 18.04+
Windows
- Windows 10
OSX
- Mojave 10.14.6+
Recommended Hardware:
Recommended specifications are: at least 16 Gb of RAM, ideally 32 Gb, and more for very large images, a CPU with at least 4 cores, preferably 16 or more, and a recent NVIDIA graphics card such as a RTX series card. Older graphics cards could work but may cause trouble or be too slow. Aydin Studio's summary page gives an overview of the strengths and weaknesses of your machine, highlighting in red and orange items that might be problematic.
Known Issues:
Here are some issues that are being actively addressed and will be addressed asap:
-
Stop button for all algorithms. For technical reasons having to do with the diversity of libraries we use, we currently cannot stop training. We are planning to solve this using subprocess spawning. For now, to stop Aydin from running, you need to close the window and/or terminate the process. We know this is very unfortunate and are determined to fix this quickly.
-
On Ubuntu and perhaps other Linux systems, high-dpi modes tend to mess with font and ui element rendering.
Road Map:
Planned features:
- Toggling between 'Advanced' and 'Basic' modes to show and hide advanced algorithms.
- Loading of denoising model and configurations (JSON) on Aydin Studio
- Pytorch backend
Contributing
Feel free to check our contributing guideline first and start discussing your new ideas and feedback with us through issues.
249 Dec 28, 2022
2.2k Dec 31, 2022
97 Dec 01, 2022
19 Nov 27, 2022
2 Nov 01, 2022
616 Jan 06, 2023
805 Jan 09, 2023
5 Mar 05, 2022
8 Aug 28, 2022
207 Dec 23, 2022
76 Feb 04, 2021
2 Feb 09, 2022
731 Jan 01, 2023
99 Dec 09, 2022
4 Feb 22, 2022
33 Jan 05, 2023
165 Nov 27, 2022
65 Jan 07, 2023
62 Dec 06, 2022
2 Feb 10, 2022