This is a work-in-progress Python-port of Mark Kittisopikul's
'Adaptive Resolution Orientation Space'
https://github.com/mkitti/AdaptiveResolutionOrientationSpace
Great work, Mark!
I stuck quite close to the original code structure and variable names, but for simplicity I excluded the possibility of inputing parameter spaces, though there a remnants of it in the code. Also, I do not plan to port the graphical user interface.
I develop and run the code in an Anaconda-environment with the additional packages 'colorcet' and 'scikit-image'.
Just download and copy the repository into the directory you're working on.
The following snippet should give you an idea about how to use the code:
from steerableAROSD import steerableAROSD
from parameters import DefaultParams
ip = DefaultParams()
package = steerableAROSD(image, ip)
The function steerableAROSD() just requires an image in form of a 2D
numpy-array and a dict() holding all the parameters. The parameters and their
default values are specified and documented in the class DefaultParams in
the parameters.py-file. The default values are the same as in
Mark Kittisopikul's original Matlab code. The returned package is a dictionary
holding intermediate output images under keys resembling figure titles of the
original code.
Ah, there will also be plenty of warnings be printed to the console, it's a bit messy at the moment, sorry for that!
This project is licensed under the GNU GPLv3 License - see the LICENSE.md file for details