CSBDeep in Fiji
This wiki page explains how you can setup Fiji/ImageJ to execute the example networks on different input data.
See the Installation Instructions for how to install and setup Fiji for the CSBDeep commands.
The plugin provides a set of CSBDeep model demonstrations which are located under Plugins > CSBDeep > Demo.
- Open an image, yours or one from our provided dataset
- Choose a suitable CSBDeep command
- Wait until all steps of the computation are completed and the result is displayed
You can also run your own network by using the plugin located under Plugins > CSBDeep > Run your network.
You can find more detailed instructions for individual commands below.
- Each CARE network should be trained with data for a specific combination of image content (e.g., nuclei, microtubules) and image corruption (camera noise, pixel size, microscope PSF, etc.). Hence, applying trained networks to images that are very dissimilar to the training data could lead to unexpected results. The pretrained networks provided via Fiji and KNIME are meant to showcase our method on the accompanying example data.
- Software for generating training data and training of CARE networks will be released soon. This will enable all users to employ CARE on their own image data.
See Fiji Command – 3D Denoising (Tribolium)
| Input | Output |
|---|---|
| 3D grayscale image | Prediction as a 3D grayscale image |
See Fiji Command – 3D Denoising (Planaria)
| Input | Output |
|---|---|
| 3D grayscale image | Prediction as a 3D grayscale image |
See Fiji Command – Surface Projection (Flywing)
| Input | Output |
|---|---|
| 3D grayscale image | Prediction as a 2D grayscale image |
See Fiji Command – Isotropic Reconstruction (Retina)
| Input | Output |
|---|---|
| 3D image with 2 channels | Prediction as a 3D image with 2 channels |
See Fiji Command – Deconvolution (Microtubules)
| Input | Output |
|---|---|
| 2D grayscale image or time series | Prediction with the same dimensions |