Release Notes
New deploy mode was implemented to export YAPiC models to DeepImageJ Bundled models
Keras model files can be converted with a YAPiC command into a DeepImageJ bundled model with all necessary metadata, example images and conversion scripts included. The model can be directly applied in DeepImageJ. There is no need of executing `DeepImageJ:Build BundledModel` inside ImageJ.
Tensorflow and Keras dependencies were removed
In previous versions, Tensorflow version 2.1. was automatically installed.
We removed the Tensorflow dependecy to allow the user to install the Tensorflow version that fits best to the specifications of their GPU hardware. Keras functions are now imported from the tensorflow backend.
Tests were added to support multiple tensorflow versions
New Tutorials were added to YAPiC website
Tutorials how to train a model with the leaf example dataset were updated. The leaf example dataset was also updated and is now compatible with Ilastik version 1.3.3. Tutorial part II was added to demonstrate deployment of YAPiC models to DeepImageJ.
Probability maps as multichannel tif images
Probability maps can are exported as multichannel images (default setting). The old way of exporting multiple single channel images is still available using the flag `--split`
Some minor bugs were resolved in yapic-io
Contributors
* Christpoh Möhl
* Hardik Doshi