I will build an offline face recognition system with python and opencv
Independent System Architect
About this Gig
I build recognition systems that identify people from a camera and run fully offline: no cloud, no monthly fees, no GPU needed.
Built and tested on a low-cost ARM Linux device (4 cores, USB webcams) at about 6 fps per camera. It names every person in view, several at once.
What you get:
Python + OpenCV detection and recognition (YuNet + SFace)
Tracking per person, so names stay stable
Web dashboard (FastAPI): live view, registration with consent, daily report, CSV export
Entrance + exit camera mode, or single-camera kiosk mode
HDMI monitor output and auto-start on boot (systemd)
Full source code and a README
Good for: attendance, door logs, visitor counting, or any project that needs to know who is in front of a camera.
Please message me before ordering with your camera, your hardware and what you want to record, so I can confirm the right package.
Expertise:
Image processing
•
Object detection
Programming language:
Python
Tools:
OpenCV
My Portfolio
FAQ
What hardware does it run on?
I built and tested it on a 4-core ARM Linux device (Armbian) with USB webcams. It also runs on a normal PC. For Raspberry Pi or other boards, message me first so I can check speed on your hardware.

