Teaching your webcam to read hand signs

What it does

Emojinator captures your webcam feed, classifies your hand gesture, and overlays the matching emoji image. It is a computer-vision toy built in Python and OpenCV that maps physical hand signs to digital reactions. The repository houses three separate iterations—V1, V2, and V3—alongside the ghosts of spin-offs like Rock Paper Scissors Lizard Spock that have since moved to other repos.

The interesting bit

The README offers a textbook definition of emojis yet stays completely silent on model architecture, dependencies, or accuracy. That earnest opacity is almost charming. What is evident is that the project evolved through three iterations, with the original version exposing its own training script and application entry point.

Key highlights

• Three distinct iterations tracked in separate folders
• Recognizes roughly a dozen hand-gesture classes mapped to PNG emoji overlays
• Spawned standalone spin-offs: Rock Paper Scissors Lizard Spock and EmojiGram
• Built with Python and OpenCV using webcam input
• 500 stars, though the codebase dates to 2018

Caveats

• The README omits model architecture, dependencies, and accuracy metrics entirely
• Differences between V1, V2, and V3 are left for the reader to diff
• Several sub-projects have been moved to other repos, leaving this one structurally fragmented

Verdict

Worth a look if you want an end-to-end OpenCV gesture-recognition toy to dissect. Skip it if you need documented training pipelines or modern framework support.