OCR, 8 alpanumeric characters (split characters only)

IN PROGRESS
Bids
1
Avg Bid (USD)
$1785
Project Budget (USD)
$30 - $5000

Project Description:
I'm looking to have an OCR written which can solve the captchas in the attached .zip file.

They are basically blue on white clean images with no noise and little skew. The real trick is separating the characters. It may help to know that the solutions are always 8 characters, but the character width can vary dramatically.

I've written OCRs myself in the past so I'm not going to ask for an unrealistic accuracy rate. I'd be happy with 2-5% depending on the $cost and the CPU load. If you can do better, that would be awesome.

The target system is linux but if you're a Windows-only programmer that's fine as well -- I'll write some kind of gateway for it. Either way I will need source code. If that's going to be a problem please mention that in your bid and we can discuss it.

Please, no subscription-based services, hidden or otherwise. This OCR will be tested on a machine without an internet connection so your app cannot pass the image off to remote typers. I'm looking for a real OCR here.

3rd party and open source components are fine.

Potential bidders will be supplied with the URL to the actual captcha; just ask.

Happy bidding!


*** Update ***

Since I can handle the OCR recognition piece of it myself, it seems easier all-around to entertain bids for the character separation/splitting portion only.

For example, given an image captcha.jpg, your app would simply need to generate letter1.jpg, letter2.jpg, ... letter8.jpg. BMP's or anything else is fine as well. The only thing is, they do need to be clean and not have significant strokes from neighboring letters. One bad letter will of course spoil the entire image.

I would expect somewhere around 20+% of the 8-character images to be split cleanly, ** or your app can just return an error code if that one looks too hard **. That percentage is negotiable.

Please submit bids for the separation portion only, and if you'd like to handle the OCR itself as well you can mention a cost for that as well in your text.

Thanks!

## Deliverables

vWorker.com requirements notice: As originally posted, this project does not have complete details. Should a dispute arise and this project go into arbitration "as is", the contract's vagueness might cause it to be interpreted against you, even though you were acting in good-faith. So for your protection, if you are interested in this project, please work-out and document the requirements onsite.

Skills required:
C Programming, Linux, Microsoft, Software Architecture, UNIX, Windows Desktop
Additional Files: vw_2012___02___23___captcha_samples_RAC_NameCryptedToProtectYourPrivacy_X2012223133138874570371205646503733028653.zip vw_2012___02___23___image01_RAC_NameCryptedToProtectYourPrivacy_X2012223133138874570401862564473318832564.jpg
About the employer:
Verified
Public Clarification Board
Bids are hidden by the project creator. Log in as the employer to view bids or to bid on this project.
You will not be able to bid on this project if you are not qualified in one of the job categories. To see your qualifications click here.


$ 1785
in 14 days