Seitenhistorie

Disclaimer: Noise2Void improves the visual quality of an image. However, it is uncertain whether or not signals remain scientifically accurate

...

for quantification (morphological quantification may be fine, intensities not)

Useful for:

• Visualization
• Intermediary step for mask generation
• Segmentation

...

1. Start Noise2Void with the icon on the desktop : N2V _OMERO Image Added

2. Open the Login dialog and insert Start by entering your credentials for the server omero-cai.hhu.de

Image Removed

3. Get ready for the training of a model for your images:

and click on "Connect"

Image Added

4. You can then select the group in which the dataset you want to process is located, and indicated the dataset ID. Load the dataset by pressing "Load dataset". You can change which image ID, Z, T or C to display by updating the field and clicking on "Set"

5. To train the network to denoise your images, go to the tab "Train N2V"

1. Patch size: Together with "Batch size", will change how much data is passed to the GPU per batch. A patch of 256x256 with a batch size of 16 works best for our workstation (the one next to the spinning disk microscope)
2. Number of training epochs: Increase this if your loss is still decreasing after 100 epochs.
3. Select a dataset on OMERO with its ID
4. Get an image to test the connexion
5. → Jump to 5. if you already have a trained model
6. Choose a name for the model
7. Select the path where to save the model
8. "Number of training epochs" and "Batch size" can be left as default (if there is a memory error, for the training, the batch size must be reduced).
9. Click on "Start training" and go for a break. This should last 10-15 min.

...