General Midway Equalization

Python implementation of general midway equalization using arbitrary number of grayscale images, .

Example runs using grayscale images avaliable in the following jypitor notebook

general_midway_equalization.ipynb

It is a requirement that all images have same dimensions.

Very short intuition behind code

Let denotes the normalized culumative histogram of a grayscale image, , and the pseudo-inverse of .

The midway equalization method can be generalized to N arbitrary number of images. Specifically,

where for and .

I refer to utils/functions.py for detailed implementation. For further describtions and alternative methods please see [1] and [2].

How to run existing code

• Step 1: Open general_midway_equalization.ipynb.
• Step 2: Load N number grayscale images and make sure they all have same dimension. Stack them on top of each others either in a list or array.
• Step 3: Run codes from utils/functions.py and plot results.

---

Try it out yourself, and if you have any questions don't hesitates to create an issue. Cheers!

References

[1] Julie Delon. Midway Image Equalization. In: "Journal of Mathematical Imaging and Vision", Springer Verlag, 21 (2) (2004), pp.119-134. DOI: 10.1023/B:JMIV.0000035178.72139.2d

[2] Thierry Guillemot and Julie Delon. Implementation of the Midway Image Equalization. In: "Image Processing On Line", 6 (2016), pp. 114-129. DOI: 10.5201/ipol.2016.140