Intel® Integrated Performance Primitives
Community support and discussions relating to developing high-performance vision, signal, security, and storage applications.
Announcements
This community is designed for sharing of public information. Please do not share Intel or third-party confidential information here.

## BGRtoLAB wrong formula specified?

Beginner
215 Views

HI

I have a question about the function BGRtoLAB

for LAB, there is a case condition specified in the Manual:

L = 903.3 * (Y/Yn)1/3. for Y/Yn 0.008856

Now, if I take the border case ,  0.008856^(1/3) is ~0.207,   ->  mult  903.3   = ~187

This value is converted to range of 0..255 by factor 2.55, so we get 476 which is way out of range.

Actually i run into this condition (Y/Yn 0.008856) for values like  R=0, G=0, B=10 which is pretty dark, L should be Close to 0 also

So what is wrong with your formulas?

thx & BR

Stefan

6 Replies
Beginner
215 Views

Hi again

found some other formulas. I think it should be without Exponent ^(1/3) like in the f(t) formulas:

L = 903.3 * (Y/Yn). for Y/Yn ≤ 0.008856

as 903 is the factor 7.787*116

Employee
215 Views

Thank you for reporting the issue,

We will investigate it.

Beginner
215 Views

Yes, have a look. the code is ok, the results are ok of the ipp, just the doku

also in the f(t) formula something is wrong, as f(t) is around ~1 you probably won´t substract 16 and multiply later the result .... . It´s all messed up

Employee
215 Views

the code is ok, the results are ok of the ipp, just the doku

What is "doku"? Besides, what does "the code is ok, the results are ok of the ipp" mean? Your program works OK with the IPP function BGRtoLab?

Is this a “theoretical” question? Or could you provide us with a test image and a sample code that can reproduce the problem that you encountered, please?

Beginner
215 Views

Yes, the code works, the results are ok

only the documentation (docu, doku) is horribly wrong (the named formulas)

BR

Employee
215 Views

OK. Got it. Thank you for reporting the issue.