I have a question for the paper.
If we do not have color image (ground truth) during inference, is it impossible to provide local/global hint (patches that the user give) for the model to colorize?
When I tried colorization on a grayscale image of my own data, the model provided hint of grayscale patches.
So if I understood the paper correctly, are the hints provided in inference stage when we only have colored image?
Hello Prof. Zhu.
I have a question for the paper. If we do not have color image (ground truth) during inference, is it impossible to provide local/global hint (patches that the user give) for the model to colorize?
When I tried colorization on a grayscale image of my own data, the model provided hint of grayscale patches.
So if I understood the paper correctly, are the hints provided in inference stage when we only have colored image?
Thank you in advanced.