Deep dive with Fiji

The story of CSBDeep

Context

Who am I?

Trained Media Computer Scientist
♥ Open Source
Research Software Engineer

                       1
                      0
                     ___
                     |1|
                     |0|
      .             .'1'.
     /       __    /   0 \     / __   /
   _/__(_/__(_/_  /_______\  _(_(_/|_/_)
  _/_       _/_   ':::::::'
 (/        (/

Deep learning for microscopy at CSBD

Deep Learning in Fiji?

Toolset for
Deep Learning in Fiji

imagej-tensorflow

Java library for dealing with TensorFlow in ImageJ
https://github.com/imagej/imagej-tensorflow

Edit › Options › TensorFlow..

CSBDeep for Fiji

Toolkit for running image to image prediction
on pretrained networks
https://imagej.net/CSBDeep/

Update Site: https://sites.imagej.net/CSBDeep/

Plugins › CSBDeep › Run your network

Integration

CSBDeep can be used as a library to run
prediction from a script / plugin
e.g. StarDist, DeepImageJ are using CSBDeep

Plugins › CSBDeep › Demo › ..

Training

CSBDeep (python)

Unsupervised training with N2V

Training

Training

N2V for Fiji

one-click solution for content aware denoising
https://github.com/juglab/N2V_fiji/

Update Site: https://sites.imagej.net/N2V/

Future work

BioImage Model Zoo

https://bioimage.io/
https://github.com/bioimage-io

?

YAML Specification
https://github.com/bioimage-io/configuration_draft

ImageJ Model Zoo Consumer
https://github.com/imagej/imagej-modelzoo/