Browse categories
Explore
Fiverr Pro
English
$
USD
I will develop an ai powered image detection and text extraction system
Dusan Dj.
About this gig
Gig Description
Need to automatically detect objects in your images or pull text out of documents, receipts, ID cards, plates, or scanned files? I build AI-powered systems that turn raw images into clean, structured data no manual work required.
What I can build for you:
- Custom object detection (products, vehicles, defects, people, plates, packaging, etc.)
- Text extraction (OCR) from images and scanned documents
- Combined pipelines: detect the object read the text inside it output clean data
- Structured results delivered as JSON, CSV, or connected to a database
- Model fine-tuning on your own images if you have a specific use case
- Integration into your existing app (API wrapper) if needed
Why work with me:
I built a complete AI detection and text-extraction pipeline as part of my Software Engineering thesis combining object detection with OCR into one working system, tested on real, messy, real-world images (not just clean demo data). I focus on systems that actually hold up outside a perfect test environment.
Tech stack: Python, YOLOv8, OpenCV, EasyOCR, PyTorch, FastAPI
Not sure which package fits your case? Message me before ordering happy to take a quick look and advise honestly, eve
Get to know Dusan Dj.
Dusan Dj.
Software Developer
- FromSerbia
- Member sinceNov 2023
- Avg. response time1 hour
- Last delivery6 months
Languages
English, Serbian, Slovak
Hi there, I'm Dušan!
I'm a Software Developer with 6+ years of experience building full-stack web applications, dashboards, and business systems. I work with Angular, Java Spring Boot, and MySQL, creating modern, scalable, and secure solutions for clients from various industries.
Throughout the years, I’ve worked on multiple projects — from internal company tools and management systems to customer-facing apps. My focus is always on clean architecture, performance, and a smooth user experience.
FAQ
Do you need my images to train a custom model?
For custom training, yes — ideally 200+ labeled images for solid accuracy. I can also help with labeling for an extra fee.
Can I get this integrated into my existing app, not just as a script?
Yes — the Premium package includes integration and a simple API (FastAPI) so it connects directly to your system.
What if my images are low quality?
I'll tell you upfront if image quality will limit accuracy — I don't promise results I can't deliver.
What kinds of objects/text can this detect?
Almost anything with enough example images — products, vehicles, documents, plates, packaging labels, etc. Message me your use case and I'll confirm it's feasible.
Do you provide the source code?
Yes, source code with clear comments is included in every package.
Can you connect this to a database?
Yes, database integration is available in the Premium package.
Do I need to know anything about AI or coding to use this?
No — you'll receive working code and clear documentation. If you want it fully integrated into your app without touching any code yourself, choose the Premium package.
