As you already know, I'm writing utility program for conversion of scanned scientific books to PDA and other devices with small screen size. There is a great need in such a program since there is no reliable and free formula recognition product.
I discovered quite simple solution - we don't need to recognize text, we can just reformat the images of pages to feet to PDA screen.
So that's all about history and motivation. Now let's talk about new version. In this version I incorporated deskew algorithm which automatically rotates image to make text horizontal. Also I reworked splitting code, and now the program is able to split colorful images (though, it uses only black-and-white data to find lines of text) and it tries to preserve under- and over- lines.
I was thinking a lot about coding style when writing this code, because I didn't have sufficient C/C++ practice last couple of years. So I will be happy to hear your suggestions on that matter.
Also I want to give sounding name to this program, so I declare some kind of contest for the best name, and I'm waiting for your proposals.
And one announcement: next version will have a GUI, so I plan to make it public. I will be happy with any kind of support.
Update:
I forgot to mention sexy example
No comments:
Post a Comment