Dear organizers,
What inputs will we have during the inference? Is it possible to use all filtered versions of the same scene to generate "unfiltered" prediction? Or only filtered HR and LR are given to predict the original image? Can we have access to the filter meta info (e.g. name) when inferencing or is it completely blind?
Thank you!
Posted by: ipg_mrc @ June 3, 2022, 1:11 p.m.It should be designed as image-to-image translation problem. Consider like you have a single filtered image and you want to recover the original version of it in inference time. Also, you may not have any external/meta data to support the inference.
Posted by: birdortyedi @ June 8, 2022, 8:54 a.m.