MediaMan Community MediaMan Forum
Pages: 1
OCR By: 0_o Date: August 27, 2009
Is it possible to implement an OCR over the images imported into MediaMan?
Right now I have to scan some books, type their names and import their picture. 3 actions.
An Optical Character Recognition ability can save me time since I could scan a book for example and just drag and drop it's picture into MediaMan which will create a new object in the directory I dragged the picture into with the OCR result as it's name.
Much faster...
Moreover, I could drag a pdf into MediaMan and have it's file name or even it's name from within the file as it's item's name.
Thank you for MediaMan!
Re: OCR By: He.Shiming Date: August 27, 2009
Hmm... I'm not sure if I understood you correctly. You are referring to "reading out book names from the cover picture". Well, I would only say that theoretically this would be possible. But common books are use artistic fonts in their titles and names, and might be colorful as well, not to mention the layout. I don't think any of the OCR solutions can read such images well, given that they are mostly designed to read black & white documents in a standard font. So I'm sorry this might not be possible currently.
Re: OCR By: JC Date: August 29, 2009
There's an iPhone app called SnapTell where you can take a picture of a DVD or book and it will do some sort of OCR on it, then looks the item up on Amazon.  So while not exactly what the guy's talking about, it could be another way of looking up items (other than scanning the barcode).
Re: OCR By: He.Shiming Date: August 30, 2009
Quote from: JC on August 29, 2009
There's an iPhone app called SnapTell where you can take a picture of a DVD or book and it will do some sort of OCR on it, then looks the item up on Amazon.  So while not exactly what the guy's talking about, it could be another way of looking up items (other than scanning the barcode).
This is actually possible, which was in the plan for the next update. The idea is simple, to read barcode out of a picture. However, this post is about scanning covers and storing in MediaMan, which, as I mentioned in some level requires professional software.
Re: OCR By: JC Date: August 30, 2009
Maybe I'm reading it differently than you are.  I was taking it along the lines of taking an image of the cover of the book, and have it read the title of the book and do a lookup on Amazon based off that (vs. the barcode).  Something I've noticed in scanning my books is that the UPC on a lot of the older books is no longer valid (seems to be reused a lot for other books).  I have to scan the ISBN that's in the front of the cover or do lookups by name to get accurate lookups.

But yeah, I don't know if that level of OCR is doable without having some sort of integration with a mature OCR application.
Re: OCR By: He.Shiming Date: August 30, 2009
Sorry for the previous inaccurate reply. The post isn't about scanning covers. It's about reading the "book title" out of a cover picture, which can be tricky for the reason I mentioned earlier in the post. Reading the barcode should be easier as I said, I'm planning on including this feature.