I need a library that can detect objects in an image (uses edge detection). This is NOT related to captchas. I am working on an MTGO bot that uses OCR and that works in any screen resolution. In order for it to port to any screen resolution my idea is to scan down narrow range on a results page (the cards that a player has can be listed in rows of text) and to find each object in that range. Then to take the lowest and highest pixel coordinates of each object to find where the row starts and ends (on the y axis) so that I can use OCR to read each line.
+3
A:
If you don't know of the OpenCV collection of examples, then they could help you in the right direction... there's also Camellia which doesn't use "edge detection" per-se but could get the results you need with a bit of work.
ozone
2009-02-02 15:23:29
Can't seem to find any relevant examples. Were there some that you had in mind specifically and can point me towards? Thanks.
Zombies
2009-02-09 20:26:02
A:
Have at look at the Lead tools imaging SDKs? Not free, but very comprehensive.
Shane MacLaughlin
2009-02-02 15:57:37
Yes I plan on using tesseract-OCR, for well, the OCR part obviously. I need a way to scan the image and find objects before splicing it up into segments for the OCR to read.
Zombies
2009-02-03 15:19:49
+2
A:
Its not cheap, but I've used the Intel Processing Primitives, and was very impressed with their performance. They work on Intel and AMD processors, as well as Windows and Linux
Steve
2009-02-02 21:30:17
It's not too bad - it's ~$300 a license. I use IPP at work and am very happy with it.
unforgiven3
2009-02-02 21:32:20
A:
Framewave based on AMD Performance Library which provides features for image processing. Check also related Fast Cross-Platform C/C++ Image Processing Libraries
mloskot
2010-01-23 23:35:35