Fix loading of TIF 16bits grayscale files

[mariehbourget]

Checklist

GitHub

• I've given this PR a concise, self-descriptive, and meaningful title
• I've linked relevant issues in the PR body
• I've applied the relevant labels to this PR
• I've assigned a reviewer

PR contents

• I've consulted ivadomed's internal developer documentation to ensure my contribution is in line with any relevant design decisions
• I've added relevant tests for my contribution
• I've updated the relevant documentation for my changes, including argparse descriptions, docstrings, and ReadTheDocs tutorial pages

Description

This PR fixes issue #1105 where TIF 16bits grayscale gave an error while loading:

• It includes a new test to make sure every combination of PNG/TIF, 8/16 bits, Grayscale/RGB/RGBA files are loaded properly.
• It requires a new set of testing images being integrated in data-testing in PR Add data_test_png_tif dataset data-testing#18.
• PR Add data_test_png_tif dataset data-testing#18 needs to be approved, merged and a new release of data-testing done to update the link for download here before merging the current PR.
• I updated some keywords in the code of the test_loader.py file as well.

Linked issues

Fixes #1105

[dyt811]

Thanks for the concise PR. I will review with the PR18 data and approve it when it passes.

[dyt811]

Looking all good. I think just need to update the source ivadomed/scripts/download_data.py for the path to the release and ensuring to use the newer release:

    "data_testing": {
        "url": ["https://github.com/ivadomed/data-testing/archive/r20220221.zip"],
        "description": "Data Used for integration/unit test in Ivadomed."},

[coveralls]

Pull Request Test Coverage Report for Build 2054396579

• 3 of 3 (100.0%) changed or added relevant lines in 1 file are covered.
• No unchanged relevant lines lost coverage.
• Overall coverage increased (+0.06%) to 79.163%

---

Totals
Change from base Build 2053772380:	0.06%
Covered Lines:	4673
Relevant Lines:	5903

---

💛 - Coveralls

[akraskovSMDB]

Hi, I think I still have a problem with 16 bit tifs. Is there something wrong with my tif?

I run exactly the same training/testing using the same set of 16 and 8bit images and predictions of the model are wrong for 16 bit images but reasonable for 8 bit images.

I have fresh install of ivadomed
pip show ivadomed
Name: ivadomed
Version: 2.9.8
Here is the link to the images (I can't upload 16 bit tif, it is more than 25 Mb limit)

Happy to share whole BIDS folder and imadomed output folder with the config file

[akraskovSMDB]

Just checked and I have the same problem with 16 bit PNG images.