Contact person: Business Client

Phone:Show

Email:Show

Location: New Delhi, India

Budget: Recommended by industry experts

Time to start: As soon as possible

Project description:
"I want to layer reliable face recognition onto the live RTSP feeds coming from the CCTV cameras already installed at our premises. The goal is simple: every time someone walks in or out, the system should spot the face, match it against a database of staff and visitors, and write an “entry” or “exit” event with a precise timestamp. Those events must be searchable by name, camera or date and viewable through a lightweight, browser-based dashboard that also lets me enrol new faces on the fly and pull daily or weekly reports.

The environment you’ll step into includes about twenty HD IP cameras streaming over our internal network and a spare on-prem server (Windows or Linux) that can be equipped with a dedicated GPU if your stack—OpenCV, FaceNet, DeepFace, TensorFlow, YOLO, AWS Rekognition or another engine—needs it. Privacy is critical, so everything should run locally; cloud components are acceptable only when they can be self-hosted.

Deliverables
• A service that consumes our camera streams, detects and recognises faces, and logs entry/exit events to a database.
• A clean web dashboard (responsive) for live viewing, search, CSV export and manual face enrolment.
• Source code, build/installation scripts and a short demo using two of our cameras that proves the accuracy and logging flow.

When you reply, include a link or video of a previous face-recognition or AI-surveillance project you built, note the frameworks you plan to use, the expected accuracy you can deliver, and any hardware you’ll need so I can approve procurement quickly. I’m ready to begin as soon as I find the right partner and would like to see a working prototype within the first milestone." (client-provided description)