>embedforge
← Back to projects

Case study

Interactive Mirror

Face-detection-driven smart mirror that overlays contextual info as you walk up to it.

PythonOpenCVFace DetectionRaspberry Pi
Detection range
0.5m – 2.5m
Wake latency
<300ms
Active-time reduction
~70%

Problem

Ambient displays behind two-way mirror glass need to detect a person's presence and approximate distance/attention without a full facial-recognition pipeline that would be overkill (and privacy-invasive) for a home device.

Stack

  • Python
  • OpenCV Haar/DNN face detector
  • Raspberry Pi 4
  • HDMI two-way mirror panel

Architecture

Interactive Mirror system architecture: camera feed, face detection with landmark and pose tracking, overlay logic selecting an effect based on pose, a compositor blending the overlay onto the frame, and the two-way mirror display output.

Outcome

Detection-triggered wake reduced the display's active-on time by roughly 70% versus always-on, cutting panel burn-in risk and standby power draw.

View on GitHub ↗