Algorithm to Draw Rectangles Around Words in Images

Greetings Overflowers,

I need an algorithm to draw rectangles around words in images of scanned documents. This algorithm needs not to be 100% accurate anyway. Though, it is preferable to require the least of human intervention during its work. No assumptions on开发者_Go百科 scanned document before hands (unknown fonts and languages, might include tables and figures)

Any ready made solutions or resources on the matter ?

Regards

most OCR engines provide this kind of information. But if you don't want to use an engine OpenCV can be used for this but it will require model building and tuning.

OCR in OpenCV http://blog.damiles.com/?p=93

That's a big step. In short this is a hard problem to solve and if you want to do it yourself it's going to take a good while to get it working. Most likely you should consider a 3rd party OCR package. Like Nuance. I've use it and can verify it has exactly this type of information and a lot more.

http://www.nuance.com/for-individuals/by-product/omnipage/index.htm