Note that we did not include the high-definition versions of the images in this submission. This is due to file size concerns. They are available online if you want to re-run that particular set yourself.
skimage numpy scipy (should be included with skimage)
This library assumes you're using Python 3, specifically a version 3.6 or greater
In order to run a specific image through the algorithm, you can issues the following command
python3 src/main.py <path/to/image.jpg>
The resulting images will be places in results//
A number of images are generated - each of the colour channels are exported separately, along with two composites. naive.png is a naive version where the channels are just overlapped. wirth.png is the result of the registration.
If you want to run a set of images, just point the script at an entire directory.
"python3 src/main.py <path/to/images/>
main.py is a wrapper that calls our library.
All functionality is located in BunchLib.py. Relevant functions that you might actually care about include:
- alignChannels - our actual alignment algorithm
- colorizeGorskiiImgWirth - wrapper that calls the above and exports channels
- colorizeGorskiiImgNaive - wrapper that creates the naive version of the image
The rest are largely helper functions and import/export functionality