It is currently 16 Apr 2024, 08:27
   
Text Size

Upscale low res card images to high res using Deep Learning

Discuss Card Scans and Other Artwork Here

Moderator: CCGHQ Admins

Re: Upscale low res card images to high res using Deep Learn

Postby austinio7116 » 22 Jan 2019, 07:37

User avatar
austinio7116
 
Posts: 451
Joined: 10 Mar 2017, 11:59
Has thanked: 47 times
Been thanked: 169 times

Re: Upscale low res card images to high res using Deep Learn

Postby AStrangerIsHere » 22 Jan 2019, 11:17

Thanks!
AStrangerIsHere
 
Posts: 164
Joined: 22 Jan 2019, 01:17
Has thanked: 45 times
Been thanked: 29 times

Re: Upscale low res card images to high res using Deep Learn

Postby Bog Wraith » 25 Jan 2019, 00:12

Thanks as always for doing this austinio. I missed seeing them for UMA and am very happy to see these beauties & in time for the release weekend festivities!

2 cards are missing from the Upscaled RNA set that are in the Forge database. They are from the Planeswalker decks. Confusing thing is that there are other cards from these decks that are not in the Forge DB. They'll probably be added along with the basic lands later on.

Here are the two .png files from The Gatherer:
Attachments
Dovin's Automaton.full.png
Ragefire.full.png
'Twas in the bogs of Cannelbrae
My mate did meet an early grave
'Twas nothing left for us to save
In the peat-filled bogs of Cannelbrae.
User avatar
Bog Wraith
Global Mod 1 (Ret)
 
Posts: 1108
Joined: 28 May 2008, 22:40
Location: Shandalar
Has thanked: 425 times
Been thanked: 153 times

Re: Upscale low res card images to high res using Deep Learn

Postby austinio7116 » 20 Apr 2019, 07:25

User avatar
austinio7116
 
Posts: 451
Joined: 10 Mar 2017, 11:59
Has thanked: 47 times
Been thanked: 169 times

Re: Upscale low res card images to high res using Deep Learn

Postby austinio7116 » 20 Apr 2019, 07:27

Angrath, Captain of Chaos.full.png
User avatar
austinio7116
 
Posts: 451
Joined: 10 Mar 2017, 11:59
Has thanked: 47 times
Been thanked: 169 times

Re: Upscale low res card images to high res using Deep Learn

Postby AStrangerIsHere » 20 Apr 2019, 14:26

Goo job! But it's kind of funny how the quality of the cards is quite high, but when you look at the name of the illustrator at the bottom of the card, and it's blurry. Any reason why?

Edit: and there's twice the first mountain.
AStrangerIsHere
 
Posts: 164
Joined: 22 Jan 2019, 01:17
Has thanked: 45 times
Been thanked: 29 times

Re: Upscale low res card images to high res using Deep Learn

Postby austinio7116 » 20 Apr 2019, 15:17

I think the artist text is just too small in the low res input images. Thanks for letting me know about the mountains, there must have been duplicates on the Wizards site.
User avatar
austinio7116
 
Posts: 451
Joined: 10 Mar 2017, 11:59
Has thanked: 47 times
Been thanked: 169 times

Re: Upscale low res card images to high res using Deep Learn

Postby Bog Wraith » 20 Apr 2019, 16:38

A problem that sometimes slips by is that the card names with an apostrophe are changed like this when up-scaled, "Liliana's Triumph.full" becomes "Liliana_s Triumph.full".

There are 45 files that I have corrected and have uploaded them for anyone who wants them to use in apps like Forge. I only use the 300dpi sized cards so those are the only one's I corrected. Just swap these in for the underscored ones:

https://www.mediafire.com/file/2z6p416x ... d.zip/file
'Twas in the bogs of Cannelbrae
My mate did meet an early grave
'Twas nothing left for us to save
In the peat-filled bogs of Cannelbrae.
User avatar
Bog Wraith
Global Mod 1 (Ret)
 
Posts: 1108
Joined: 28 May 2008, 22:40
Location: Shandalar
Has thanked: 425 times
Been thanked: 153 times

Re: Upscale low res card images to high res using Deep Learn

Postby austinio7116 » 20 Apr 2019, 18:38

It's not the upscaling that messes up the name, it's Google drive. I keep forgetting, it's best if I zip the files and upload the full zips.
User avatar
austinio7116
 
Posts: 451
Joined: 10 Mar 2017, 11:59
Has thanked: 47 times
Been thanked: 169 times

Re: Upscale low res card images to high res using Deep Learn

Postby austinio7116 » 20 Apr 2019, 18:41

When I get a free couple of minutes I'll upload some zips.
User avatar
austinio7116
 
Posts: 451
Joined: 10 Mar 2017, 11:59
Has thanked: 47 times
Been thanked: 169 times

Re: Upscale low res card images to high res using Deep Learn

Postby austinio7116 » 20 Apr 2019, 18:43

austinio7116 wrote:It's not the upscaling that messes up the name, it's Google drive. I keep forgetting, it's best if I zip the files and upload the full zips.
In fact it is only when you download them the filenames are changed, in Google drive they are correct.
User avatar
austinio7116
 
Posts: 451
Joined: 10 Mar 2017, 11:59
Has thanked: 47 times
Been thanked: 169 times

Re: Upscale low res card images to high res using Deep Learn

Postby austinio7116 » 20 Apr 2019, 18:46

Zips are uploading now for anyone that wants them to avoid the "_" in filenames issue
User avatar
austinio7116
 
Posts: 451
Joined: 10 Mar 2017, 11:59
Has thanked: 47 times
Been thanked: 169 times

Re: Upscale low res card images to high res using Deep Learn

Postby austinio7116 » 31 May 2019, 22:45

Here are the results for Modern Horizons:

https://drive.google.com/open?id=1nyRO7 ... hJqHh2sQKK

Archmage's Charm.full.jpg
User avatar
austinio7116
 
Posts: 451
Joined: 10 Mar 2017, 11:59
Has thanked: 47 times
Been thanked: 169 times

Re: Upscale low res card images to high res using Deep Learn

Postby austinio7116 » 01 Jun 2019, 06:25

Tokens added now too.
User avatar
austinio7116
 
Posts: 451
Joined: 10 Mar 2017, 11:59
Has thanked: 47 times
Been thanked: 169 times

Re: Upscale low res card images to high res using Deep Learn

Postby AStrangerIsHere » 01 Jun 2019, 11:39

Austinio, my man!!! As always, thx!

Btw, there's a little error on Wrenn and Six > "Wrennd Six".
AStrangerIsHere
 
Posts: 164
Joined: 22 Jan 2019, 01:17
Has thanked: 45 times
Been thanked: 29 times

PreviousNext

Return to Pictures

Who is online

Users browsing this forum: No registered users and 36 guests


Who is online

In total there are 36 users online :: 0 registered, 0 hidden and 36 guests (based on users active over the past 10 minutes)
Most users ever online was 4143 on 23 Jan 2024, 08:21

Users browsing this forum: No registered users and 36 guests

Login Form