simple_native_vision

A high-performance Flutter plugin for computer vision and on-device AI tasks, providing MediaPipe Face Landmarks, ONNX Runtime document corner detection, TFLite object detection, and active liveness analysis.

Features

• Active Liveness Detection:
  • Real-time face tracking and pose estimation (pitch, yaw, roll).
  • State-based challenge validation (look straight, turn left, turn right, smile, etc.).
  • MediaPipe Face Mesh backend on Android & iOS.
• Document Corner Detection:
  • Edge and corner inference powered by ONNX Runtime (onnxruntime-android / onnxruntime-objc).
• Object Detection:
  • Mobile target detection using TensorFlow Lite.
• Lightweight Modular Architecture:
  • Standalone vision library — keep your main app lean if you only need standard camera/device capabilities (simple_native).

Getting Started

Add simple_native_vision to your pubspec.yaml:

dependencies:
  simple_native_vision: ^0.0.1

Usage

import 'package:simple_native_vision/simple_native_vision.dart';

// Example: Configure active liveness thresholds
final config = LivenessThresholdConfig(
  yawThreshold: 20.0,
  pitchThreshold: 15.0,
);

// Perform AI / Vision tasks using SimpleAI

License

BSD 3-Clause License. See LICENSE for details.