Browse categories
Explore
Fiverr Pro
English
$
USD
I will build computer vision and image processing solutions using opencv and python
Rohit J.
About this gig
Need to detect, track, count, or recognize something from images or video? I build production-grade computer vision systems that actually work in real conditions.
I'm Rohit, a Computer Vision engineer at IIT Mandi. I've deployed CV pipelines in real robotics competitions including a QR-based inventory detection system achieving 92% accuracy under variable lighting at Inter IIT Tech Meet 2025, where my team won Bronze among 23 IIT teams.
This isn't tutorial-level work. I build systems that run.
What I can build:
Object detection & tracking (YOLO, OpenCV DNN, custom models)
Barcode & QR code detection/decoding
Face detection & recognition
Color & shape detection
OCR text extraction from images/documents
Image classification (binary or multi-class)
Motion detection & background subtraction
Defect detection / quality inspection systems
Real-time video processing (webcam, RTSP streams)
Deliverables: Python source code (clean, documented), requirements file, sample output images/video, setup instructions.
Tech stack: OpenCV · Python · PyTorch · YOLO (v8/v11) · NumPy · scikit-learn
Message me before ordering with what you need to detect and your input source.
Get to know Rohit J.
Rohit J.
Building systems that actually work
- FromIndia
- Member sinceDec 2025
- Avg. response time1 hour
- Last delivery3 days
Languages
Hindi, English
B.Tech student at IIT Mandi (Robotics & Autonomous Systems), Technical Co-Coordinator of the Robotics Club — mentor of 50+ students, Bronze Medal at Inter IIT Tech Meet 14.0.
I build web apps, dashboards, automations, and MVPs fast — using Python, C++, ROS2, React, Tailwind, and Supabase/Vercel.
Custom CRM, data pipeline, AI automation, robotics simulation — if you can describe it, I'll deliver clean, deployable code with fast turnaround and revisions.
My Portfolio
FAQ
I only have a small dataset. Can you still help?
Yes. I can use data augmentation and transfer learning to maximize accuracy with limited data.
What do I need to provide?
Sample images or video showing what to detect, and a description of the desired output. More examples = better accuracy.
Do you train custom models?
Yes — Standard and Premium include training on your data. You need labeled images (200+ per class for basic tasks, 500+ for good accuracy).
Can you work with live camera feeds?
Yes. I build real-time solutions for webcams, IP cameras, or RTSP streams.
What accuracy can I expect?
For well-lit, consistent scenes: 85-95%+. For challenging conditions I give an honest estimate upfront before you order.
I only have a small dataset. Can you help?
Yes. I use data augmentation and transfer learning to maximize accuracy with limited data.
