
Automated form-submission workflow
- or -
Post a project like this29
£363(approx. $502)
- Posted:
- Proposals: 12
- Remote
- #4469038
- OPPORTUNITY
- Open for Proposals
PPH's TOP Notch Website & Mobile App Developer & Designer(9+ yrs) ✔ Wordpress ✔ Shopify ✔ OpenCart ✔ Laravel ✔ PHP ✔ React Native ✔ Android ✔ iOS ✔ HTML/CSS✔Javascript/jQuery✔Responsive Design✔ASP.net




Expert Web Developer - N8N, Wordpress, Shopify, Opencart, Laravel, Vue, React, PHP

Microsoft Certified AI Developer | .NET / .NET Core Architect & Full-Stack Developer

1181845611903143197371115255352744138471735115009776299797559836857951716668710261563
Description
Experience Level: Expert
I’m building an automated form-submission workflow and need a C# component that reliably deciphers text-based CAPTCHAs that use custom fonts. The images are not wildly distorted, but the lettering differs from standard OCR sets, so off-the-shelf libraries struggle.
Here’s what I’m after: a reusable class or small DLL that accepts a bitmap/byte array of the CAPTCHA and returns the decoded string quickly enough for live form submissions. You’re free to use Tesseract, IronOCR, Emgu CV, or a lightweight CNN you train yourself—as long as the final code runs under .NET 6, compiles cleanly, and requires no paid dependencies.
Please include a minimal console project that demonstrates the call pattern and shows typical accuracy on a handful of sample images I’ll supply. If you train a model, ship the trained weights alongside the code so I can retrain later if the provider tweaks the font. Target accuracy is 90 %+ on the sample set; feel free to suggest preprocessing tricks to get there.
Once delivered I’ll drop the reader into my service and test against live traffic, so clarity, clean error handling, and straightforward licensing are a must.
Here’s what I’m after: a reusable class or small DLL that accepts a bitmap/byte array of the CAPTCHA and returns the decoded string quickly enough for live form submissions. You’re free to use Tesseract, IronOCR, Emgu CV, or a lightweight CNN you train yourself—as long as the final code runs under .NET 6, compiles cleanly, and requires no paid dependencies.
Please include a minimal console project that demonstrates the call pattern and shows typical accuracy on a handful of sample images I’ll supply. If you train a model, ship the trained weights alongside the code so I can retrain later if the provider tweaks the font. Target accuracy is 90 %+ on the sample set; feel free to suggest preprocessing tricks to get there.
Once delivered I’ll drop the reader into my service and test against live traffic, so clarity, clean error handling, and straightforward licensing are a must.
James A.
0% (0)Projects Completed
-
Freelancers worked with
-
Projects awarded
0%
Last project
30 Jan 2026
United Kingdom
New Proposal
Login to your account and send a proposal now to get this project.
Log inClarification Board Ask a Question
-
There are no clarification messages.
We collect cookies to enable the proper functioning and security of our website, and to enhance your experience. By clicking on 'Accept All Cookies', you consent to the use of these cookies. You can change your 'Cookies Settings' at any time. For more information, please read ourCookie Policy
Cookie Settings
Accept All Cookies