I will build yolo object detection, tracking and counting computer vision apps
About this Gig
Need to detect, count or track people, vehicles or products in video? I build computer vision systems with YOLO and OpenCV that turn camera footage into usable numbers.
My final-year project is a retail analytics pipeline: YOLO detection, Re-ID matching across cameras, homography to floor coordinates, and footfall-to-sales funnel analysis.
Services:
- Object detection on images or video (YOLOv8/v11)
- Custom model training on your own labeled data
- Dataset labeling setup (CVAT / Roboflow) and cleaning
- Multi-object tracking (ByteTrack, BoT-SORT) and Re-ID
- Line crossing, zone counting, dwell time and heatmaps
- Export to ONNX/TensorRT, FastAPI endpoint or Streamlit demo
You get the trained weights, clean Python code, an accuracy report (mAP, precision, recall) and instructions to run it.
Message me with a sample image or short clip before ordering.
APIs:
Google Cloud Vision API
Programming language:
Python
•
SQL
•
Colab
Tools:
Jupyter Notebook
•
OpenCV
•
CVAT
•
Colab
•
PyTorch
Frameworks:
Scikit-learn
•
Keras
•
PyTorch
FAQ
Do I need labeled data?
For a custom model, yes. I can set up labeling or label small datasets for an added fee.
Can it run in real time?
Usually yes on a GPU; I'll tell you the expected FPS for your hardware.
Do you work with live CCTV/RTSP streams?
Yes, RTSP and webcam inputs are supported.
Is face recognition included?
No. I focus on detection, tracking and counting, not identifying individuals.

