Hi everyone,
I need some help analyzing dorsal fur coloration in a set of hundreds of images I received. In the past, I’ve used micatoolbox for this type of analysis with RAW images, where I had the grey reflectance values, making it straightforward. However, the images I’m working with now are in JPG format and use Kodak color control patches.
I understand that if I can’t create a custom non-linear model, I might have to use sRGB, though I know that comes with risks. My biggest challenge is finding reliable standard reflectance values for these Kodak patches, which I haven’t been able to locate online.
Are these images still usable for analysis, or is there a way to extract accurate information from them despite these issues?
Here’s an example of one of the images: https://imgur.com/a/SwkxZxT
Thanks in advance for any guidance!
Personally I would say that we shouldn’t write off image datasets that aren’t calibrated. We just need to be careful about how we present the results, and be extremely cautious of any systematic bias in sampling that might affect the images.
But worse case scenario you could just put the images through as sRGB. I don’t know about the kodak colour control patches (or how to find out about them). But if there’s a grey patch on them, and it’s consistent among images then as long as you choose a sensible grey value and use this for the whole image set, any error in absolute value will just scale linearly among images. You wouldn’t be able to compare the absolute reflectance values among other datasets, but for testing stuff within this dataset it should be fine.