open_iA 2022.06

New or vastly extended tools:

• CoSi
• ImNDT
• XVRA

Other improvements:

• new click+drag interactions in slicer: ROI transfer function or window/level adjustment
• raw file parameter guessing from file name
• added filter: multiply images
• FeatureScout: Improvements and fixes for class loading and mean object handling; a FeatureScout project now stores defined classes.
• and many more improvements and fixes (slicer tooltips, chart data export, job list layout, dark theme, ...)

Full changelog: 2021.08...2022.06

Portable, pre-built binaries for Windows are attached below (AI and Astra modules are only available in the larger "with-CUDA" version). Just extract the .zip file, and execute open_iA.exe.

A linux version is available in our flatpak repository.

Should you encounter any issues, or should you have any suggestions for improvements, let us know in the issue tracker!

SHA512 checksums:

• open_iA_2022.06.zip: c1c671a2763c0ae98d1d70be8664651a23970772b8ba2afe0b89e03f87992313e0854e686e1751bd71eabdb0f95eaff540c14e0b18bcddda79e91b3b82dbf3f0
• open_iA_with-CUDA_2022.06.zip: 88a93d39563aad048bfd90bea77b8b4af505f07fb858f04115e56ba8ecad06843cbecdbb96d01f1d1a84b73ea908d92d7ceb6dbe62ff4c12edc4c4c638b551bc

Additional resources:

• To run onnx models in our AI module, you'll have to download CuDNN (version 8.0.5.39 for CUDA 10.2) and put the cudnn*.dll's either in the same directory as open_iA or in another directory that is in the PATH environment variable and before all other potential cudnn installation folders in PATH when open_iA is run.
• Core API documentation for 2022.06
• As test datasets, use those from the 2020.09 release.