LYiHub

LYiHub / AR-Mahjong-Assistant-preview

Public

一个基于向听算法的AR眼镜端麻将助手

558
144
100% credibility
Found Feb 17, 2026 at 102 stars 5x -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
Python
AI Summary

An augmented reality system that uses smart glasses to capture Mahjong hands, analyze them locally with AI, and overlay optimal discard or call suggestions.

How It Works

1
🔍 Discover AR Mahjong Coach

You hear about a cool helper that uses smart glasses to give real-time tips during Mahjong games.

2
💻 Set Up on Your Computer

Download and launch the coaching program on your home computer with a simple one-click start.

3
🥽 Connect Your Smart Glasses

Link your AR glasses to the computer over your home Wi-Fi network so they can talk to each other.

4
🎮 Begin Your Mahjong Session

Put on the glasses, tap to start a new game, and get ready to play with invisible coaching.

5
📸 Snap Your Hand Tiles

Tap or slide on the glasses to capture a quick photo of your tiles and any called sets.

6
💡 View Best Move Advice

Watch smart suggestions for the perfect tile to discard or calls to make appear right on your glasses lens.

🏆 Play Smarter Mahjong

With instant tips on every turn, you make better decisions, learn faster, and enjoy winning more often!

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 102 to 558 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 AR-Mahjong-Assistant-preview?

This Python project turns RayNeo AR glasses into a real-time mahjong assistant, capturing hand photos via the glasses' camera, analyzing them on a local server with computer vision and shanten algorithms, and overlaying optimal discard or call suggestions directly on the AR display. It solves the pain of mental math during intense games by providing instant strategy like best cuts, tenpai waits, or meld opportunities, all running offline for privacy and speed. Voice input lets you describe opponent discards or calls to refine "dead tile" tracking.

Why is it gaining traction?

Unlike cloud-dependent apps, this preview stays fully local with ONNX YOLO for tile detection and Docker one-click setup, delivering sub-second latency even on modest hardware. Developers dig the voice-to-LLM integration for natural game state updates, plus a web debug tool for tuning detection thresholds on custom lighting. At 90 stars, it's hooking mahjong AI tinkerers who want AR hardware hacks without vendor lock-in.

Who should use this?

Mahjong players with RayNeo X3 glasses competing in casual or competitive riichi sessions, needing discreet edge without apps. AR/Android devs prototyping wearable AI for tabletop games. Chinese mahjong fans tweaking local LLMs like Qwen for voice commands during live play.

Verdict

Grab it if you have the glasses and play mahjong—solid docs and Docker make setup painless, but 90 stars and 1.0% credibility signal prototype risks like finicky recognition. Test on your table first; it's a clever local assistant worth forking for polish.

(198 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.