Speed Benchmarks for morphological operations

[ogencoglu]

Would be useful to have an understanding of run-time performance of the most common morphological operations such as

• erosion
• dilation
• thinning
• closing
• opening
• skeletonization
  etc. with respect to other popular libraries such as:
• opencv
• scikit-image
• scipy.ndimage
• fastmorph
• imops

[crisluengo]

Sure, that would be nice.

Here’s some comparisons with scikit-image: https://github.com/grlee77/uskimage-demo?tab=readme-ov-file#2d-results

At the bottom of this page there is a comparison of the dilation with OpenCV: https://www.crisluengo.net/archives/1140/

Here’s a comparison of the median filter with OpenCV and scikit-image: https://www.crisluengo.net/archives/1138/

I’ve also done some comparisons of the dilation with MATLAB’s Image Processing Toolbox: https://www.crisluengo.net/archives/885/

It would be great if you or someone else did a more thorough comparison across libraries and algorithms!