Hello, you could optimize it for comics, that it recognizes separately the text of the balloons and the drawings so that it applies a different method for each case.
Like this old scan:
I use RealESRNET and Photoshop to optain this, but can be better I thing:
Hello, you could optimize it for comics, that it recognizes separately the text of the balloons and the drawings so that it applies a different method for each case.
Like this old scan:![Old Scan](https://user-images.githubusercontent.com/719690/132073339-c20efd1c-77c3-41a3-a172-01141f7979ba.jpg)
I use RealESRNET and Photoshop to optain this, but can be better I thing:![RealESRNET](https://user-images.githubusercontent.com/719690/132073669-25212bc3-8f4b-408a-b1ee-30413ef6d948.jpg)
I make a comparison here vs Gigapixel: http://paconaranjo.eu5.org/sliders/SuperResolution/CRG/index.html
Thanks!