I’ve been using NLP for a while now but this is the first time the algorithm got confused with an image. It’s interpreting the data wrong unless I include a part of the film border.
It’s been whitebalanced properly before.
First image is when its cropped properly with wrong colors and the second is with some part of the film border included and more correct colors. I can send you the raw file if you need it for debugging. Is it because there is no true black in the image otherwise?