Vision Camera Pose Landmarks Plugin

A Vision Camera Frame Processor Plugin for real-time human pose detection using MediaPipe Pose Landmarker. This plugin includes the (pose_landmarker_full.task) model bundled with the package.

Features

• 🚀 Real-time pose detection (30+ FPS on modern devices)
• 🎯 33 body landmarks with 3D coordinates (x, y, z)
• 📊 Visibility and presence scores for each landmark
• 🔄 Optimized for video streams with temporal tracking
• 📱 Works on both Android and iOS devices

Platform Compatibility

| Android Device | Android Emulator | iOS Device | iOS Simulator | Web | | -------------- | ---------------- | ---------- | ------------- | --- | | ✅ | ✅ | ✅ | ❌ | ❌ |

Installation

npm install vision-camera-pose-landmarks-plugin
# or
yarn add vision-camera-pose-landmarks-plugin
# or
pnpm add vision-camera-pose-landmarks-plugin

Usage

Basic Example

import { usePoseLandmarksPlugin } from 'vision-camera-pose-landmarks-plugin';
import { useFrameProcessor } from 'react-native-vision-camera';

function App() {
  const { detectPoseLandmarks } = usePoseLandmarksPlugin();

  const frameProcessor = useFrameProcessor((frame) => {
    'worklet';
    const landmarks = detectPoseLandmarks(frame);

    // Access specific landmarks
    if (landmarks.nose) {
      console.log('Nose position:', landmarks.nose);
    }
  }, []);

  return <Camera device={device} isActive frameProcessor={frameProcessor} />;
}

Credits

• Built with MediaPipe
• Powered by Vision Camera