Skip to content

ELECTRON_PERFORMANCE.md

Latest commit

 

History

History
99 lines (72 loc) · 2.17 KB

ELECTRON_PERFORMANCE.md

File metadata and controls

99 lines (72 loc) · 2.17 KB

Electron Performance Optimization Guide

Tips untuk meningkatkan performa Electron app agar tidak lag.

🚀 Quick Fixes (Sudah Diterapkan)

Optimasi berikut sudah otomatis aktif di Electron:

✅ Lower Resolution: 640x480 (vs 1280x720 di web) ✅ Single Hand Detection: Hanya 1 tangan (vs 2 di web) ✅ Disabled Connections: Tidak render skeleton lines ✅ Adaptive Frame Skipping: Skip frame jika processing lambat ✅ GPU Disabled: Menghindari GPU crash ✅ Smaller Buffer: History buffer lebih kecil

🎯 Manual Optimizations

1. Disable Visualization (Paling Efektif!)

Klik tombol "Toggle Visualization" untuk mematikan overlay.

Atau edit src/js/config/electronConfig.js:

visualization: {
  enabled: false  // Matikan semua visualization
}

Impact: +10-15 FPS

2. Lower Camera Resolution

Edit src/js/config/electronConfig.js:

camera: {
  resolution: {
    width: 320,   // Sangat rendah tapi cepat
    height: 240
  }
}

Impact: +5-10 FPS

3. Reduce Target FPS

performance: {
  targetFPS: 20  // Lower = less CPU
}

Impact: Lebih smooth, less CPU

4. Increase Detection Confidence

mediapipe: {
  minDetectionConfidence: 0.7,  // Higher = less processing
  minTrackingConfidence: 0.7
}

Impact: +2-5 FPS, tapi kurang sensitif

5. Disable Specific Gestures

Edit src/js/config/gestureConfig.js:

// Disable gestures yang tidak dipakai
pinch: {
  enabled: false
},
push: {
  enabled: false
}

Impact: +1-3 FPS per gesture

📊 Performance Comparison

Setting Web Mode Electron Default Electron Optimized
Resolution 1280x720 640x480 320x240
Max Hands 2 1 1
Connections Yes No No
Target FPS 30 30 20
Expected FPS 30-60 20-30 25-40

🔧 Advanced Optimizations

1. Use Offline Model

Download model untuk menghindari network latency:

# Download model
curl -o src/models/hand_landmarker.task https://storage.googleapis.com/mediapipe-models/hand_landmarker/hand_landmarker/float16/1/hand_landmarker.task