Hello!
Our Beta release 0.1.0 is ready to go.
Here's where the code and instructions on how to use it are located: https://github.com/ccipd/collageradiomics
This is the Git repository that houses our code and issue tracker. There are instructions provided showing how to install our Python module and how to to explore the examples within our Docker container. Our module is packaged in a standard format and is simple to use for those familiar with the ecosystem, but please feel free to email if you have any questions about Python or pip.
As per the instructions, our examples are housed here: https://github.com/ccipd/collageradiomics/tree/master/jupyter/examples
The instructions also mention this, but I'd like to note that the examples require Docker to be installed before running. This ensures that the development experience is standardized and prevents errors related to individual environment setup.
We have four examples:
1. Demonstration of running CoLlAGe on a rectangular mask defined in code.
2. Demonstration of running CoLlAGe on a non-rectangular mask passed in as an image.
3. Demonstration of running CoLlAGe on multiple images and multiple masks.
4. Demonstration of running CoLlAGe on a single feature.
The output of the code is primarily an image with the CoLlAGe algorithm executed on the passed in mask. There are secondary outputs of the raw Haralick features, and the features overlayed onto the image in code.
Both examples provide a demonstration of how you can take this output and display it in an easy to read way:
CoLlAGe Feature 10
I have attached a full preview in PDF format of what the non-rectangular example looks like after it executes.
Should anyone have feedback or encounter issues, please have direct them to our issue tracker and we will respond quickly: https://github.com/ccipd/collageradiomics/issues
As you can see in the issue tracker, there are some things we'd like to add (hosted documentation, unit test suite, multiple modules, and more!) as we go further in the project, but the core is solidified and is ready for alpha testers to use. We welcome any questions or suggestions.
It's been fun getting to this point-- we're excited to move forward into the Slicer integration, and we look forward to seeing what students and researchers develop with your groundbreaking work.
Developer Contact Info:
Nathan Hillyer
tothtechhillyer.me
Senior Engineer - TothTech - Custom AI Solutions Specializing in Medical Imagery
TothTech.ai