tjcoding / enhanced-image-colour-transfer-2 Goto Github PK

I had this program in both the most recent version of chrome and firefox. The issue is as shown in the console: The response containing CT_noLUT.wasm, initiated by a request from CT_noLUT.js, does not have a content-type header. When I manually specified "Content-Type: application/wasm", as suggested by the console output, it worked. You can use chrome devtools to locally override response headers.

Another thing you may want to note: at first, I was using jpg images that were around 5mb, and the application would crash. It seems the images need to be a smaller file size for the application to succeed. I don't know what is the root cause.

*edit: Just wanted to thank you guys for making both the software and the documentation available. It's been a big help in improving my understanding of color, color spaces, etc, and I wanted to say how appreciative I was. I've built my understanding of the relevant science/math based on the incredibly generous contributions of so many people to open source projects, free textbooks, and in general the relatively thankless enterprise of sharing knowledge. So thanks.

Originally posted by @metamaterialsuit in #2 (comment)

This would be amazing if it could have a python implementation.
It would make it a lot more accessible to people and further the integration opportunities.

When generating the output image, the process hangs.

it would be interesting if it were possible to convert a large number of source images that are in the same folder using a batch

Hi, I am not familiar with C++, and I am not sure if this is the case. In line 147 of the file main.cpp

// Implement image refinements where a change is specified.
SaturationProcessing(targetf, savedtf,
PercentSaturationShift/100.0);
targetf=FullShading(targetf, savedtf, sourcef, ExtraShading,
PercentShadingShift/100.0);
targetf=FinalAdjustment(targetf,savedtf,
PercentTint/100.0,
PercentModified/100.0);

Should the sentence be

targetf=SaturationProcessing(targetf, savedtf,
PercentSaturationShift/100.0);

Maybe Saturationprocessing is not necessary, so it is written like this?

After reading the Reinhard paper and looking into the Ruderman color space, I have noticed something that appears to be a contradiction.
In your work, you say that correlation can be used to enhance the accuracy of the color transfer algorithm, but it seems like Reinhard et all specifically wanted to avoid correlation beween the color channels.
Am I interpreting this correctly?