2018-12-06 - I2K 2018 Dresden Hackathon

Revision as of 08:50, 17 December 2018 by Maarzt (talk | contribs)

From Monday, December 3, 2018 through Friday, December 7, 2018, the Max Planck Institute of Molecular Cell Biology and Genetics hosts ~10 developers at the Center for Systems Biology in Dresden, Germany for a hackathon to develop KNIME Image Processing and ImageJ2/Fiji core infrastructure and plugins.

Gitter channel: https://gitter.im/juglab/Hack1812

Topics / Technical Discussions

Labkit as Labeling Editor for KNIME Image Processing

  • Node for editing labels in KNIME based on Labkit was created.
  • Work is not entirely finish. We plan to finish this work during the Ostrava Hackathon.
  • Labkit was extended, and is now able to handle multiple images.
  • A labeling datatype to store labels, and track modifications was developed.

Simple ImageJ API

  • Discussion between: Deborah, Robert, Curtis, Matthias
  • How can we make the scripting in ImageJ2 easier?
  • Setup a table, of scripting examples in IJ1 and IJ2. Improve the IJ2 API where ever it's obviously required.
  • Write a SimpleIJ gateway, that provides really simple methods for the most frequent tasks in ImageJ.
  • A ImageJ2 script recorder could improve the IJ2 API, as it clearly shows where the API is to verbos.

Big Data Server

  • Discussion Tobias, Gabriella and Matthias about a mechanism for secure user authentication in Big Data Server (and ImageJ Server).
  • It's needed by biologists, that want to share there data with specific users.
  • ImageJ Server would also benefit, if the access to the API could be authenticated in a similar way.

Octtree Like, Compressed, Fast, Bitmaps for Imglib2

  • Tobias develops a data structure to store bitmap. For iteration, efficient pixels access and little memory foot print at the same time.
  • Such data structures can be used to better express labels in imglib2. Even instance segmentations that consist of millions of labels, can be accessed efficiently if a cell image like cache is used.