In the paper, I see you proposed a DnCNN-3 model which can either denoise a picture or perform super resolution on a picture. However if I want to denoise a picutre and perform super resolution on it at the same time, is this model capable of this task? I trained a model by myself but the result is very poor if the input is both noisy and bicubic magnified. Could you offer me some advice on it?
In the paper, I see you proposed a DnCNN-3 model which can either denoise a picture or perform super resolution on a picture. However if I want to denoise a picutre and perform super resolution on it at the same time, is this model capable of this task? I trained a model by myself but the result is very poor if the input is both noisy and bicubic magnified. Could you offer me some advice on it?