MyanNet - Burmese Digit Recognition

A lightweight CNN with TensorFlow.js for real-time Burmese digit recognition. Live demo with drawing pad and camera input.

Test Accuracy: 99.49% | Model Size: 24.18 KB | Parameters: 10,634

Quick Start - Deploy to Vercel

Method 1: GitHub (Recommended)

1. Initialize Git repository:

   git init
   git add .
   git commit -m "Initial commit: MyanNet deployment"

2. Push to GitHub:

   • Create a new repository on github.com
   • Copy the remote URL
   • Run:

     git remote add origin https://github.com/YOUR-USERNAME/REPO-NAME.git
     git branch -M main
     git push -u origin main

3. Deploy on Vercel:

   • Go to vercel.com/new
   • Import your GitHub repository
   • Click Deploy
   • Your app will be live in ~1 minute!

---

Method 2: Vercel CLI (Fastest)

1. Install Vercel CLI:

   npm install -g vercel

2. Deploy:

   vercel

   • Link to your Vercel account
   • Accept defaults
   • Done! Your URL will appear in the terminal

---

Method 3: Drag & Drop

• Visit vercel.com
• Drag and drop the entire folder onto the dashboard
• Instant deployment

---

Local Testing

python -m http.server 8000

Open http://localhost:8000 in your browser.

---

Architecture

• Block 1: Conv2D(64,3×3) + BN + ReLU + MaxPool
• Block 2: DWConv(3×3) + BN + PW(1×1,64) + BN + MaxPool
• Head: GAP → Dropout(0.194) → Dense(64) → Softmax(10)

---

Model Info

• Framework: TensorFlow.js (browser-based)
• Weights: Real trained weights (base64 embedded)
• No server required — runs entirely in your browser
• Works offline after initial page load

---

Features

✅ Live Drawing Pad — Draw digits and get instant predictions
✅ Camera Input — Real-time digit detection
✅ Research Details — Architecture, results, cross-validation
✅ Edge Deployment — Optimized for embedded systems
✅ Mobile Friendly — Responsive design

---

Citation

IEEE Research Demo · BHDD Dataset · Marwadi University · 2026