Browse categories
Explore
Fiverr Pro
English
$
USD
I will build a real time object detection system using opencv python
AI ML, Computer Vision, Web Scraping, FastAPI Backend, Flutter Apps
I am a professional Software Engineer specializing in building AI systems and production-grade software that solve real business problems — not just demos.
My work includes computer vision pipelines ...
About this Gig
Most object detection projects fail not because of the model but because of poor integration, slow inference, and no real-world testing.
I build production-grade object detection systems using OpenCV and Python that run in real time, on your actual hardware, with your actual data.
Custom-trained detection models your classes, your dataset
Real-time video detection RTSP streams, webcam, video files
Multi-object tracking with DeepSORT and ByteTrack
ONNX export for CPU deployment no GPU required
FastAPI endpoint wrapper for direct integration
️ Autonomous driving ADAS system 95%+ accuracy Hong Kong client
️ Real estate property analysis ResNet + FastAPI Dubai client
️ People counting, vehicle detection, defect inspection all shipped
If you have a dataset, I will train on it. If you do not, I will guide you on exactly what you need.
Message me before ordering. Describe your detection task and I will tell you what accuracy to expect!
My Portfolio
FAQ
What if I don't have an annotated dataset?
I can annotate your raw images as part of the project. For Standard and Premium packages, mention this when ordering and I'll include annotation in the scope.
How many images do I need for good accuracy?
Minimum 100-200 images per class for usable results. 500+ per class for production-grade accuracy. I'll advise on your specific case, just describe the task.
Which detection model do you use?
Primarily YOLO for the best speed accuracy balance. I also work with newer and older versions depending on your hardware and accuracy requirements. Tell me your constraints.
Can this run on CPU without a GPU?
Yes, I export to ONNX format which runs on CPU efficiently. Inference speed drops compared to GPU but the model works fully. Good for edge deployment and local servers.
What do I receive on delivery?
Trained model weights, ONNX export, inference Python script, evaluation report with accuracy metrics, and a README with usage instructions.
