Browse categories
Explore
Fiverr Pro
English
$
USD
I will do object detection and object tracking using yolov8 and opencv
About this Gig
Your camera sees. Mine understands.
A regular model detects. A good one tracks, counts, and recognizes behavior in real time, on your hardware. That's what I build.
What you get:
- Custom YOLO model trained on your exact dataset not some generic pretrained weights
- Real-time multi-object tracking with DeepSORT, ByteTrack, or Norfair objects don't just get detected, they get followed
- Automated counting people, vehicles, industrial items, anything with a consistent shape
- Speed optimization for your hardware whether it's a Jetson Nano or a cloud GPU, I tune it to run properly
- Action & behavior recognition detect specific movements, not just presence
- Full delivery: Python script + trained weights (.pt) + documentation you can actually follow
Built with: Python, PyTorch, TensorFlow, OpenCV, YOLOv8/v10/v11, MediaPipe, DeepStream
Why me? Pretrained weights on COCO are not a custom solution. If your use case is specific factory floor, parking lot, retail store, sports analysis you need a model trained on your data. That's exactly what I do, and I optimize it for the hardware you actually have.
Message me before ordering.
FAQ
Do I need to provide the dataset for custom model training?
Yes, if you want to detect specific objects not found in standard models, you'll need to provide the images. However, I can also assist with data annotation (labeling) or help you find open-source datasets if needed.
Can your system track objects across different camera frames?
Yes, I use advanced tracking algorithms like DeepSORT or ByteTrack which assign unique IDs to each object. This allows the system to follow and count specific objects even if they move or briefly disappear from the frame.
What are the hardware requirements for real-time tracking?
For high FPS (frames per second) and real-time performance, an NVIDIA GPU is recommended. However, for less demanding tasks, I can optimize the models (like YOLOv8-tiny) to run efficiently on a standard CPU or Raspberry Pi.
Will I get the source code and the trained model weights?
Absolutely! Upon completion, I will provide the full Python source code, the trained weights file (e.g., .pt or .onnx), and instructions on how to set up the environment and run the script.
