To my best knowledge, other methods train their networks using vimeo_triplet which contains three consecutive frames instead of using vimeo_setuplet (used in video super-resolution). Why do you provide the "vimeo_setuplet" dataset for testing? Is it a fair comparison with the existing approaches?
tarun005
commented
3 years ago
To my best knowledge, other methods train their networks using vimeo_triplet which contains three consecutive frames instead of using vimeo_setuplet (used in video super-resolution). Why do you provide the "vimeo_setuplet" dataset for testing? Is it a fair comparison with the existing approaches?