Intellesis Denoising

This module allows you to train and use deep learning based models for the denoising of images. This method can be applied to any type of image and any dimensions and it is not dedicated to a special field of application. It should be used before applying processing functions that modify the pixel values to that image.

Denoising is an operation to reduce noise in an image, in case of Intellesis Denoising with the help of deep learning methods. In general, there are different ways to train a denoising model. Intellesis Denoising uses the approach called Noise2Void (N2V), which requires only a noisy input image for the training of a model and can thus be trained directly on the data that should be denoised. To give a simplistic explanation, this N2V method replaces pixels by masked pixel data randomly selected within a certain window/surrounding area. With this approach, the model is then trained to reconstruct the original pixels and to discard the implicit noise in the image. For detailed information on Noise2Void , see the paper "Noise2Void - Learning Denoising from Single Noisy Images" written by Alexander Krull, Tim-Oliver Buchhol and Florian Jug, see also https://arxiv.org/abs/1811.10980.

Licensing and Functionalities of Intellesis Denoising

Some functionality of Intellesis Denoising is generally available, but the full functionality requires the AI Toolkit license.

Basic functionality

The general available functionality includes:

Importing and exporting models.
Managing the models, including renaming and deleting.
Running a model with the Intellesis Denoising tool on the Image Processing workbench or in OAD (if you have licensed the Developer Toolkit).

Licensed functionality

If you have licensed this functionality and activated it under Maintenance > Toolkit Manager, the following additional functionality is available:

Creating and training a new model.
Retraining an existing model.

Creating and Training an Intellesis Denoising Model

You have licensed the AI Toolkit and activated it under Maintenance > Toolkit Manager.
You are in Manage Templates.
If required, you have pre-processed your image(s) with the Whitening tool, see Whitening Tool.

For Show, select Intellesis Denoising Models.
In the top right tool bar, click .
The user interface for training opens.
In the top left corner, click Import Images. To open an image from the archive, click Import from Archive.
A file browser or the archive browser opens.
Select the image for training from the file system and click Open. When loading the image from the archive, select the image and click OK.
For images with multiple channels, a dialog opens to select the channel for denoising.
Select the channel you want to use for denoising and click OK.
The image is displayed in the list. Note that all imported images are added to your model.
If you have imported an image from the archive, it is copied to the location defined under Maintenance > General Options > Intellesis.
Select the image from the list.
The image is displayed in the left image container.
In the parameter section, set the Number of Epochs and adapt the more advanced parameters, if necessary.
Click Train and Denoise.
Your model is trained based on the settings and the prediction is displayed in the right image container.
If you are satisfied with the result, click Finish.
All changes are saved, and the training window closes.

You have successfully created and trained a model for denoising. You can now use it to denoise your images with the Intellesis Denoising tool, see Using an Intellesis Denoising Model for Image Processing.

Importing an Intellesis Denoising Model

You have a trained denoising model available for import.

Under Manage Templates > Show, select Intellesis Denoising Models.
In the top right tool bar, click .
A file browser opens.
In the file browser, select the model file from the file system.
Click Import.

You have imported a model and it is available in the list of denoising models. You can now use it to denoise your images, see Using an Intellesis Denoising Model for Image Processing.

Exporting an Intellesis Denoising Model

You have a denoising model available.
You are in Manage Templates.

For Show, select Intellesis Denoising Models.
All denoising models are displayed.
Right-click the model you want to export and select Export.
A file browser opens.
Select the file location and click OK.

The model is exported. It contains the trained routine to denoise an image and is not intended for the training process anymore.

Using an Intellesis Denoising Model for Image Processing

Denoising Several Channels in Job Mode

In Job Mode, you have the possibility to add the Intellesis Denoising tool multiple times if you want to denoise several but not all channels of your image, e.g. if you want to denoise three out of five channels, you can add the tool three times.

You have a denoising model available.
You are in Free Mode or Job Mode and have added the Intellesis Denoising tool.
You have opened the image(s) for which you want to use the model.
If required, you have pre-processed your image(s) with the Whitening tool, see Whitening Tool.

In the Intellesis Denoising tool, click .
The Open Template dialog opens.
Select the Model you want to use and click OK.
In Free Mode, click Apply.
The objects in your image are denoised based on the trained model.