Priyanshu Zihowa View code on GitHub (opens in a new tab)

Computer vision · Public safety · Privacy-first

Smart CCTV Safety Monitor

Detects people, follows each one with an ID, and raises alerts when someone enters a restricted zone, loiters, or a crowd forms. Everything runs in your browser: the video never leaves your device.

Loading the detection model…

Drag on the video to draw a restricted zone. People are checked by where their feet are.

How it works

  1. 1

    Detect

    A pretrained EfficientDet-Lite0 model (MediaPipe) finds people in every frame and returns a box and a confidence score for each.

  2. 2

    Track

    A tracker I wrote matches each box to the same person in the next frame using overlap (IoU) and a constant-velocity prediction, so every person keeps one ID.

  3. 3

    Apply rules

    Each person's foot point is tested against the zone polygon (ray casting). Entering triggers an intrusion alert, staying triggers loitering, and too many people triggers a crowd alert.

  4. 4

    Respect privacy

    No video is uploaded. Turn on privacy blur to pixelate everyone, so the system can count and alert without showing faces.