lovro-i/CIEDE2000

incorrect values

Opened this issue · 5 comments

nschloe commented

For some LAB pairs, this implementation gives incorrect results. For example

(50.0, 2.49, -0.001)
(50.0, -2.49, 0.0009)

Reference (http://www2.ece.rochester.edu/~gsharma/ciede2000/)

7.1792

Result from this library:

7.219468035072701
shameempk commented

Feel free to use https://github.com/shameempk/pyciede2000.
README is inspired from this project. Thanks @lovro-i .

  • 👍1
nschloe commented

I already implemented it myself in coloria. This issue report is just a heads-up that the function computes values incorrectly which may lead users on the wrong path.

shameempk commented

@nschloe That's a cool library.

denshigomi commented

Guess I should have read the issues before using this library. I just spent two hours debugging my program before discovering some values are off. I came here to post this issue, but it's already been reported. I see the same values as nschloe, which don't match the test data provided to verify the algorithm is implemented correctly.

  • 👍1
nschloe commented

This repository should probably be flagged or removed as it leads people on the wrong path, but the author doesn't reply and afaik there currently is no way of pushing this on a github level.