+ Computer Vision

Face Mask Detection

Detects masked and unmasked faces in images, video and live webcam streams with an OpenCV DNN face detector and a MobileNetV2 classifier.

View source on GitHub

Streamlit web dashboard with three input options: image upload, video file upload, and live webcam processing

Overview

A computer vision app that finds faces with an OpenCV DNN (Caffe) detector and classifies mask / no mask with a MobileNetV2 model. It supports images, video files and live webcam or CCTV streams, draws colour-coded boxes with confidence, flags close face pairs with a simple distancing heuristic, and ships with both a CLI and a Streamlit dashboard.

Highlights

  • Image, video and live webcam/CCTV inputs
  • OpenCV DNN face detection + MobileNetV2 classifier
  • Streamlit dashboard and command-line tools

Tech stack

  • Python
  • TensorFlow
  • Keras
  • MobileNetV2
  • OpenCV
  • Streamlit