CutOut

Overview

Flutter package for CutOut image segmentation implementation that enables the utilization of C++ code for image processing. This package provides a bridge between Flutter and native C++ image segmentation models, offering efficient and high-performance image segmentation capabilities in Flutter applications. The integration of C++ code allows developers to leverage existing computer vision algorithms while maintaining the cross-platform benefits of Flutter. Currently, supports iOS and Android.

Key Features

• Supports latest OpenCV 4.10.0
• Supports ONNXRuntime 1.15.1 (depends on onnxruntime_flutter 1.4.1)
• Enables conversion and utilization of Python code from research or POC projects to C++
• Supports major segmentation models including U2Net and Segment Anything(SAM)

Getting Started

dependencies:
  cutout: x.y.z

Build for iOS and Android packages

• iOS

  cd ${YOUR_APPLICATION}/ios && pod install

• Android

  cd ${YOUR_APPLICATION}/android && ./gradlew clean

Download models

You can use example models or download them models assets. The models in both locations are identical.

Update OpenCV version

Check out set_opencv.sh file. (ref. flutter_native_opencv)

Usage Example

Check out example directory.

Licenses

This package is licensed under the Apache License, Version 2.0. This package includes the following third-party models:

• U-2-Net - Apache-2.0
• Segment Anything - Apache-2.0