[WD-17] Reading Text in an image using Tesseract

Hi All,

I created a sample windev project to read text from an image. There are many opensource libraries are available. I used Tesseract for read text.

I created a windev.dll for use tesseract and included in windev project.

Its working good. The result is not 100% accurate.

You can download and check sample project from here

Regards,

Shijo S Philip



Edited 1 time(s). Last edit at 07/16/2018 12:07PM by Shijo S Philip.

HI Philip

I have been looking for something like this for ages.
Is this dll available for use or for us to help you test?

Thanks
Noel

Hi Shijo

re: Acurracy problem

A colleague of mine used the following technique to 'solve' the accuracy' problem: he interfaced THREE different recognition engines and is running them against each document. He then compares the result. When they are the same, he knows he's at 100%...

When he has two engines giving the same result, he THINKS he has 100%, but may allow a user verification.

And when he has 3 differents results, he knows that the original document is messed up, and FORCES a user review/correction.

Best regards

Hi Noel Tanti,

You can freely use this in any where.
Tesseract is opensource only.I just made a dll in visual studio for use this tesseract library in windev.

For any help on this you can contact me.

Regards,

Shijo S Philip



Edited 1 time(s). Last edit at 07/16/2018 05:48PM by Shijo S Philip.

Hi Fabrice,

That idea sounds good. But almost all recognition tools are giving same results.

In case of accuracy we can use some image processing tools to make images more clear after that we can transfer to Tesseract. Then it will give more accurate results.

Actually i did this project in python and i used open CV for image processing. i was curious to implement this in WinDev. So i created a dll in Visual studio and used in WinDev.


Regards

Shijo S Philip

to bad link is dead