113 - Histogram equalization and CLAHE

If the image histogram is confined only to a small region (low contrast images), histogram equalization can be used to stretch the histogram to include all ranges. But, this type of stretching may not result in ideal results and gives too bright and too dark regions in the image. This can be especially very bad for images with large intensity variations.

Contrast limited adaptive histogram equalization (CLAHE)
Regular histogram equalization uses global contrast of the image. This results in too bright and too dark regions as the histogram stretches and is not confined to specific region.

Adaptive histogram equalization divides the image into small tiles and within
each tile the histogram is equalized. Tile size is typically 8x8. If the image contains noise, it gets amplified during this process. Therefore,
contrast limiting is applied to limit the contrast below a specific limit. Bilinear interpolation is performed between tile borders.

This tutorial demonstrates the use of histogram equalization and CLAHE in Python to enhance low contrast images.

[moustafashomer]

Hi, I'm working as a machine learning engineer from Egypt, and our company works with images, I have never found such effective and useful materials as your videos, thank you and may god bless you.

[johnmathew]

I like your style of touching the basics when explaining advanced topics.

[ashwinramakrishnan]

Thanks for taking the time and effort to do this. The examples and presentation were great! Really helped with the work I was doing.

[feranichristy]

Hi, i currently on trying CLAHE, but i really don't understand about how to calculate cliplimit and implement it on the histogram of sub-image. Can anyone explain?

[maheshpatel]

Hello Sreenivas sir, thank you very much for detailed explanation... Great Work !!!

[gautierreynes]

Thanks a lot, this was very informative!!

[hebayousif]

Thank you for that video. I have a question regarding histopathological image: which is better to apply CLAHE or stain normalization?
Also, should we do this CLAHE method and then do pixel normalization to convert them to fit [-1, 1] or [0, 1]?

[Athulyank-Aami]

informative video.. sir one doubt, , for classifying road puddle: when i give road with shadows, my deep learning model mispredict it as road puddle.. can i use histogram equalisation to solve this?

[VidishArvind]

can u tell how we will do it if we parallelize this algorithm.

[mkproduction]

After opening GitHub many folders appears, kindly give the link of exact folder link for this code

[ernestiu]

How do you determine the clip value for CLAHE? Is it arbitrary?

[williamusuga]

First of all thanks for the videos.
what about if I need to work with a gray image?, how can I transform it in a LAB space

[boushraalshehri]

how can applay it if deal with more than one image ?