What's new
Photoshop Gurus Forum

Welcome to Photoshop Gurus forum. Register a free account today to become a member! It's completely free. Once signed in, you'll enjoy an ad-free experience and be able to participate on this site by adding your own topics and posts, as well as connect with other members through your own private inbox!

Digitizing scanned cards/tokens to individual images with transparent backgrounds


Soul Reaver

New Member
Messages
2
Likes
0
I figure there must be a good way to do this, but so far I haven't found out a way. Here's my situation:

I've scanned a whole bunch of playing cards on a flat-bed scanner. This has given me a number of digitial images, each of which consists of 9 cards on a white background.

What I'd like to do is make ONE image per card. Each of those images should have the exact same dimensions - filling the width and height of each card - and the background (the cards have rounded corners so you'd see a bit of background on a rectangular canvas) should be transparent. There should be no white left at the edges of the card in the digital image.

I'm a bit stuck as to how to do this however. I'm a real stickler for accuracy, and that makes this tough.

I'm guessing one way to do this would be to set up a template canvas of the right size and make some sort of mask or something that sets the rounded corners to be transparent, but I'm not sure if this is the best method (not to mention I'm not even 100% sure how I'd go about doing this accurately).

If I did the above, I assume I'd base the canvas on one of the existing cards first (since they're all supposed to be the same size/shape), but even that is tough - is there a generally-accepted method of getting rid of the white background space on a scan that doesn't involve laboriously selecting/deselecting with the lassoo tool? I've tried using colour range selection and the magic wand but both tend to be inaccurate (unless I'm doing something wrong?)

And then part of the problem is that the cards aren't all aligned the same on the scan (some are a bit crooked etc) - if I have to manually rotate each card and check by eye if it looks right it'll likely drive me insane well before I'm done. I'd need some way to 'snap' the cards into the canvas/mask automatically (or least get it very close). Is there some tool or method to do this in a time-saving manner?

Does anyone have any suggestions or tutorials for me to look at?
 
Last edited:

MrToM

Guru
Messages
3,595
Likes
3,321
Personally I'd have just drawn each card from scratch as a vector if 'accuracy' is of major importance, even a high PPI scan will still result in antialiasing.

I'm not sure what you mean by 'innacurate' when using the 'magic wand' or 'color range'.....both are very good at what they do even if the results are not what you want.....neither of them, nor any 'tool' for that matter, is going to do the job on its own but a combination of tools may give you a more 'accurate' end result.

As for the images on each card well yes, a mask would probably be the best way to go.

If you group the 9 images into a folder you can then apply a mask to the folder, which would have the 9 spaces ready for each image in it.
If you create the 9 'spaces' in the mask by way of a vector shape then you will remove the need to 'straighten' the cards afterwards as you could position and transform each image to fit accordingly.

This way all the cards will be of the same size, the same distance from the edge of the document, aligned to each other and square.

A possible layer stack would be something like this...

cards_A_01.png

You'll notice that each image has been converted to a smart object before transforming.
Although my images are the same the structure allows for different images.
If the image is the same for each card you can just do one and then copy it, rasterise it and duplicate a further 7 times...no need for multiple duplicates of a smart object.

'make ONE image per card' could mean either.....1 image for all 9 or 9 unique images.

Regards.
MrTom.
 

Soul Reaver

New Member
Messages
2
Likes
0
Thanks Mr Tom.

Unfortunately I don't have the option of making the cards from scratch - the whole point is to digitize the components of a physical card game. A scan is the best I can do to start with.

When I say 'inaccurate' I mean that they are unable to select the exact outside of the card only. When using colour range, the antialiasing of the image means it can't be used to accurately cut where the 'edge' of the card is - it ends up with semi-transparent areas, and selecting parts inside the card and the like. And the magic wand has similar problems (being unable to differentiate between the card and the small shadow it leaves on the white background, for example). I'm not surprised at this, but I thought that eliminating a white background from scanned photos/cards/other items must be a fairly common task for experienced photoshop users so I was wondering if anyone knew any tricks to make it easier (well, easier than doing it completely manually for every scanned item).

Each card is different unfortunately, and I have about 100 of them. While I scanned 9 per canvas initially (much easier than scanning 1 at a time), I'd like to end up with 1 card per image file when I'm done with this work - so about 100 different files.

You mentioned using a 'Vector Shape' to create the mask. I'm a beginner at Photoshop (I use CS5 by the way, I forgot to mention that) - does the creation of Vector shapes in order to make a mask involve the sort of stuff covered in the following tutorial?: http://www.photoshopessentials.com/basics/shapes/vectors-paths-pixels/ ?

I've never used Smart Objects before but I've read up on them a bit now that you've mentioned them - I'm guessing the use of 'smart objects' is to allow me to transform the file freely without loss of detail/fidelity until I'm happy with the final transform/position?
 

Top