Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

CSBDeep resources overview #14

Draft
wants to merge 5 commits into
base: master
Choose a base branch
from
Draft

CSBDeep resources overview #14

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
5 commits
Select commit Hold shift + click to select a range
8c49141
adding a structure for giving an overview of csbdeep resources
frauzufall Mar 10, 2020
bb26268
howtos: add credits, gallery, videos, FAQ
frauzufall Mar 10, 2020
7ce9ca4
howtos: improve FAQ
frauzufall Mar 10, 2020
8b9871e
adding csbdeep and stardist resources, N2V FAQ (incomplete)
frauzufall Mar 11, 2020
a2596ae
howtos: fix layout
frauzufall Mar 11, 2020
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
203 changes: 203 additions & 0 deletions docs/howto/csbdeep.html
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,203 @@
<html>
<head>
<link href="style.css" media="all" rel="stylesheet"/>
</head>
<body>
<h1>CSBDeep</h1>
<div class="csbdeep">
<div>
<h2>CSBDeep for users</h2>
Deep learning solutions based on CSBDeep:
<div class="block">
<h3>CARE networks</h3>
<h4>Related publications / credits</h4>
<div class="block">
<p>
<strong>Please see the paper in <a href="http://dx.doi.org/10.1038/s41592-018-0216-7">Nature Methods</a>.</strong>
(Preprint on <a href="https://biorxiv.org/content/early/2018/07/03/236463">bioRxiv</a>)
</p>
<p>
Supplementary material can be downloaded <a href="https://www.biorxiv.org/highwire/filestream/109407/field_highwire_adjunct_files/0/236463-1.pdf">here</a>.
</p>
<h5>Authors and Contributors</h5>
<p>
Martin Weigert<sup>1,2,*</sup>,
Uwe Schmidt<sup>1,2</sup>,
Tobias Boothe<sup>2</sup>,
Andreas Müller<sup>8,9,10</sup>,
Alexandr Dibrov<sup>1,2</sup>,
Akanksha Jain<sup>2</sup>,
Benjamin Wilhelm<sup>1,6</sup>,
Deborah Schmidt<sup>1</sup>,
Coleman Broaddus<sup>1,2</sup>,
Siân Culley<sup>4,5</sup>,
Mauricio Rocha-Martins<sup>1,2</sup>,
Fabián Segovia-Miranda<sup>2</sup>,
Caren Norden<sup>2</sup>,
Ricardo Henriques<sup>4,5</sup>,
Marino Zerial<sup>1,2</sup>,
Michele Solimena<sup>2,8,9,10</sup>,
Jochen Rink<sup>2</sup>,
Pavel Tomancak<sup>2</sup>,
Loic Royer<sup>1,2,7,*</sup>,
Florian Jug<sup>1,2,*</sup>
&amp; Eugene W. Myers<sup>1,2,3</sup>
<br><br>
<sup>1</sup> Center for Systems Biology Dresden (CSBD), Dresden, Germany<br>
<sup>2</sup> Max-Planck Institute of Molecular Cell Biology and Genetics, Dresden, Germany<br>
<sup>3</sup> Department of Computer Science, Technical University Dresden<br>
<sup>4</sup> MRC Laboratory for Molecular Cell Biology, University College London, London, UK<br>
<sup>5</sup> The Francis Crick Institute, London, UK<br>
<sup>6</sup> University of Konstanz, Konstanz, Germany<br>
<sup>7</sup> CZ Biohub, San Francisco, USA<br>
<sup>8</sup> Molecular Diabetology, University Hospital and Faculty of Medicine Carl Gustav Carus, TU Dresden, Dresden, Germany<br>
<sup>9</sup> Paul Langerhans Institute Dresden (PLID) of the Helmholtz Center Munich at the University Hospital Carl Gustav Carus and Faculty of Medicine of the TU Dresden, Dresden, Germany<br>
<sup>10</sup> German Center for Diabetes Research (DZD e.V.), Neuherberg, Germany<br>
<sup>*</sup> Co-corresponding authors.
</p>
</div>
<h4>Acknowledgements</h4>
<div class="block">
The authors want to thank Philipp Keller (Janelia) who provided Drosophila data.
We thank Suzanne Eaton (MPI-CBG), Franz Gruber and Romina Piscitello for sharing the expertise in fly imaging and providing fly lines. We thank Anke Sönmez for cell culture work.
We thank Marija Matejcic (MPI-CBG) for generating and sharing the LAP2B transgenic line Tg(bactin:eGFP-LAP2B). We thank Benoit Lombardot from the Scientific Computing Facility (MPI-CBG).
We thank the following Services and Facilities of the MPI-CBG for their support: Computer Department, Light Microscopy Facility (LMF) and Fish Facility.
This work was supported by the German Federal Ministry of Research and Education (BMBF) under the codes 031L0102 (de.NBI) and 031L0044 (Sysbio II).
M.S. was supported by the German Center for Diabetes Research (DZD e.V.).
R.H. and S.C. was supported grants from the UK BBSRC (BB/M022374/1; BB/P027431/1; BB/R000697/1), UK MRC (MR/K015826/1) and Wellcome Trust (203276/Z/16/Z).
</div>
<h4><a href="">Gallery</a></h4>
<h4><a href="">Videos</a></h4>
<h4>Source code</h4>
<div class="block">
<h5><a href="https://github.com/CSBDeep/CSBDeep">CSBDeep in Python</a></h5>
<h5><a href="https://github.com/CSBDeep/CSBDeep">CSBDeep in Java / Fiji</a></h5>
</div>
<h4>How to use the CARE networks in Python</h4>
<div class="block">
<h5><a href="https://csbdeep.bioimagecomputing.com/doc/install.html">Installation</a></h5>
</div>
<div class="block">
<h5><a href="https://csbdeep.bioimagecomputing.com/doc/models.html">Model overview</a></h5>
</div>
<div class="block">
<h5><a href="https://csbdeep.bioimagecomputing.com/doc/datagen.html">Training data generation</a></h5>
</div>
<div class="block">
<h5><a href="https://csbdeep.bioimagecomputing.com/doc/training.html">Training</a></h5>
</div>
<div class="block">
<h5><a href="https://csbdeep.bioimagecomputing.com/doc/prediction.html">Prediction</a></h5>
</div>
<h4>How to export trained CARE networks from Python for Fiji and KNIME</h4>
<div class="block">
Call this method to export after training to export a model as a ZIP which is compatible to <a href="">running prediction in Fiji</a>:
<pre><code>
model.export_TF()
</code></pre>
</div>
<h4>How to use CARE networks in Fiji</h4>
<div class="block">
<h5>Installation</h5>
<ol>
<li>Open the updater and enable update site <code>CSBDeep</code></li>
</ol>
</div>
<div class="block">
<h5>Prediction</h5>
</div>
<h4>How to use CARE networks in KNIME</h4>
<div class="block">
<h5>Installation</h5>
</div>
<div class="block">
<h5>Prediction</h5>
</div>
</div>
<div class="block">
<h3>StarDist</h3>
<h4><a href="https://github.com/mpicbg-csbd/stardist#how-to-cite">Related publications / credits</a></h4>
<!-- <h4>Gallery</h4>-->
<!-- <h4>Videos</h4>-->
<h4>Source code</h4>
<div class="block">
<h5><a href="https://github.com/mpicbg-csbd/stardist">StarDist in Python</a></h5>
<h5><a href="https://github.com/mpicbg-csbd/stardist-imagej">StarDist in Java / Fiji</a></h5>
</div>
<h4>How to use the StarDist in Python</h4>
<div class="block">
<h5><a href="https://github.com/mpicbg-csbd/stardist#installation">Installation</a></h5>
</div>
<div class="block">
<h5><a href="https://github.com/mpicbg-csbd/stardist#annotating-images">Data annotation</a></h5>
</div>
<div class="block">
<h5><a href="https://github.com/mpicbg-csbd/stardist#usage">Training & prediction</a></h5>
</div>
<h4>How to export trained StarDist model from Python for Fiji</h4>
<div class="block">
Call this method to export after training to export a model as a ZIP which is compatible to <a href="">running prediction in Fiji</a>:
<pre><code>
model.export_TF()
</code></pre>
</div>
<h4><a href="https://imagej.net/StarDist">How to use StarDist in Fiji</a></h4>
</div>
<div class="block">
<h3>N2V</h3>
<h4>Related publications / credits</h4>
<h4>Acknowledgements</h4>
<h4>Gallery</h4>
<h4>Videos</h4>
<h4>How to use N2V in Python</h4>
<div class="block">
<h5>Installation</h5>
</div>
<div class="block">
<h5>Training</h5>
</div>
<div class="block">
<h5>Prediction</h5>
</div>
<h4>How to export trained N2V model from Python for Fiji</h4>
<h4>How to use N2V in Fiji</h4>
<div class="block">
<h5>Installation</h5>
</div>
<div class="block">
<h5>Training</h5>
</div>
<div class="block">
<h5>Prediction</h5>
</div>
<h4>FAQ</h4>
<div class="block">
<h5>How long do I have to train?</h5>
Longer. 100 sepochs with 300 steps each for example.
<h5>How much data do I need for training?</h5>
Don't go much smaller than 5000K pixels, e.g. 2000x3000 or 1000x1000x5, ... The more the merrier! If you use the Fiji plugin, you place many of them in the same folder and run the "train on folder" command pointing to this folder. You can use the same folder for training and validation, it will split up the data automatically and use 90% for training and 10% for validation.
<h5>What about SEM / TEM / CMOS?</h5>
<h5>Do training and test data need to be of the same dimensions?</h5>
No. You can train on bigger images and run the prediction on smaller images. The training data is internally split up into pieces and only a random subset of each batch is fed into the network. You also train on stacks, it'll split them up again into batches internally.
</div>
</div>
</div>
<div>
<h2>FAQ</h2>
<div class="block">
<h3>GPU support</h3>
<h4>GPU support in Python (Windows, Linux)</h4>
<h4>GPU support in Java (Windows)</h4>
<h4>GPU support in Java (Linux)</h4>
</div>
<h2>CSBDeep for developers</h2>
<div class="block">
<h3>How to use CSBDeep in Python</h3>
</div>
<div class="block">
<h3>How to use CSBDeep in Java</h3>
</div>
</div>
</div>
</body>
</html>
25 changes: 25 additions & 0 deletions docs/howto/style.css
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,25 @@
html, body {
background: #e1e1e3;
}

.csbdeep {
display: flex;
flex-direction: column;
}

.csbdeep > *{
margin: 10px;
padding: 10px 20px;
}

.block {
background: #f1f1f3;
padding: 10px;
margin: 10px;
}

.block .block {
margin: 5px;
background: white;
display: block;
}