Back to Projects

Meme Matcher - Real-time Facial Expression to Meme Matching

A real-time computer vision application that matches facial expressions and hand gestures to iconic internet memes using MediaPipe's AI-powered face and hand detection.

Problem Statement

Computer vision applications for facial expression recognition are typically limited to basic emotion classification. There is a gap for creative, interactive applications that combine face and hand gesture detection in real-time for entertainment purposes.

Methodology

Leveraged MediaPipe's face mesh and hand landmark detection to capture real-time facial expressions and hand gestures via webcam. Developed a matching algorithm that maps detected expression vectors to a curated database of iconic memes. Built an interactive GUI with Tkinter for live preview and meme overlay.

Results

The application performs real-time expression matching at 30+ FPS with accurate detection of key facial landmarks and hand gestures. Users can interactively see their meme match update in real-time as they change expressions.

Tools & Technologies

PythonNumPyPathlibTkinterOpenCVMediaPipe