Valentinetemi

Real-time ASL sign language detector using MediaPipe hand landmarks + XGBoost — 98.43% accuracy, skin-tone agnostic, no CNN

13
4
100% credibility
Found Mar 23, 2026 at 13 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Jupyter Notebook
AI Summary

A webcam-based tool that recognizes American Sign Language alphabet letters in real-time by watching hand shapes.

How It Works

1
🔍 Discover the sign tool

You hear about a handy tool that recognizes sign language letters using your computer's camera right away.

2
📥 Get the ready files

You grab the simple files from the sharing page to your computer, like downloading a fun app.

3
💻 Open the viewer

You launch the viewer program, and it wakes up your camera to watch your hands.

4
👋 Sign a letter

You hold up your hand to make an 'A' or any letter, and it spots the shape perfectly every time.

5
See instant results

The screen lights up with the letter name, feeling magical as it gets most signs right on the first try.

🎉 Sign freely

Now you can practice or chat using signs, with the tool helping translate to words smoothly.

Sign up to see the full architecture

4 more

Sign Up Free

Star Growth

See how this repo grew from 13 to 13 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is asl-sign-detection-mediapipe-rf?

This Python project runs real-time ASL alphabet recognition straight from your webcam, spotting letters A-Z plus space and delete at 98.43% accuracy. Built with MediaPipe hand tracking and XGBoost classification, it delivers a skin-tone agnostic real-time ASL interpreter that sidesteps slow CNNs. You get instant webcam inference for fingerspelling, perfect as a proxy for languages like Nigerian Sign Language.

Why is it gaining traction?

It crushes alternatives with lightweight real-time detection—no raw images needed, just robust hand landmarks for consistent performance across skin tones and lighting. The 98.43% accuracy on 87,000 images hooks devs chasing real-time ASL recognition without GPU hassles. Plus, its normalization makes it scale-invariant, ideal for real-time ASL translation prototypes.

Who should use this?

Accessibility devs building real-time ASL translators or dashboards. Indie hackers prototyping webcam-based sign interpreters for deaf community tools. Researchers adapting it for NSL or other real-time gesture projects needing quick, high-accuracy starts.

Verdict

Solid proof-of-concept for real-time ASL with excellent docs and a trained model ready to run, but 13 stars and 1.0% credibility signal early maturity—prototype with it, don't ship yet. Worth forking if accessibility is your jam.

(178 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.