Cool OCR Solution Tesseract

[Cory]

While working on a project that required OCR someone suggested Tesseract and it worked surprisingly well. I didn't end up using it but it caught my eye for future macros. It's free, command line, super small, and requires no installation (IE portable). These combine to make a perfect utility for MEP. It's not very powerful or accurate on poor quality scans or handwriting but for things like crisp screen shots it works brilliantly.

 

My next planned application for it is to assist user tagging scan files. All they need to do is define a rectangle on screen where a date string or other such bit of info and using other graphics utilities I get a screen shot of that location and save it to a TIFF file. Then just hit it with Tesseract and it outputs a text file. Suck that back into MEP and use it to tag or name the file. Thus it avoids users typing.

 

Anyway I thought I would pass it along to you guys. If you do go to install it beware there is little documentation. Just know to install the Windows EXE and also install the language data. I didn't realize one needed to do the language data separately and it took me a bit to sort that out.

[perkins]

Glad you found it somewhat useful.

 

I never got around to using it much myself but I may have something in mind for it.