reusable_ml_detector 1.0.0

A reusable ML image detection widget for Flutter using Google ML Kit. Create multi-domain detectors (pets, plants, food, recycling) with simple configuration.

reusable_ml_detector

A reusable Flutter widget for ML image detection using Google ML Kit. Create specialized detectors for different domains (pets, plants, food, recycling) with simple configuration instead of duplicating code.

🎯 Features

• True Reusability: One widget, multiple domains through configuration
• ML Kit Integration: Google ML Kit Image Labeling for accurate detection
• Smart Filtering: Domain-specific label filtering and confidence thresholds
• State Management: Robust state handling (idle, loading, success, error)
• Consistent UI: Animated result cards with confidence indicators
• Easy to Use: 3-line setup for a complete detector

📦 Installation

Add this to your pubspec.yaml:

dependencies:
  reusable_ml_detector: ^1.0.0

Then run:

flutter pub get

🚀 Quick Start

1. Create a configuration

import 'package:reusable_ml_detector/reusable_ml_detector.dart';

final petConfig = DetectionConfig(
  title: 'Pet Breed Detector',
  instructions: 'Take a photo of your pet',
  confidenceThreshold: 0.6,
  relevantLabels: ['dog', 'cat', 'puppy', 'kitten', 'golden retriever', 'labrador'],
  primaryColor: Colors.orange,
);

2. Use the widget

ReusableMLDetector(
  config: petConfig,
)

That's it! You now have a fully functional ML detector.

📚 Use Cases

The same component works for multiple domains:

Pet Breed Detection

DetectionConfig(
  title: 'Pet Detector',
  confidenceThreshold: 0.6,
  relevantLabels: ['dog', 'cat', 'puppy', 'kitten'],
  primaryColor: Colors.orange,
)

Plant Species Identification

DetectionConfig(
  title: 'Plant Identifier',
  confidenceThreshold: 0.5,
  relevantLabels: ['plant', 'flower', 'tree', 'leaf'],
  primaryColor: Colors.green,
)

Food Recognition

DetectionConfig(
  title: 'Food Analyzer',
  confidenceThreshold: 0.55,
  relevantLabels: ['food', 'fruit', 'vegetable', 'dish'],
  primaryColor: Colors.red,
)

Recycling Helper

DetectionConfig(
  title: 'Recycle Helper',
  confidenceThreshold: 0.5,
  relevantLabels: ['plastic', 'metal', 'glass', 'paper', 'cardboard'],
  primaryColor: Colors.blue,
)

🏗️ Architecture

The component follows a three-layer architecture:

1. Configuration Layer (DetectionConfig)

   • Defines domain-specific parameters
   • Manages thresholds and filters
   • Customizes UI appearance

2. Detection Layer (ReusableMLDetector)

   • Handles camera/gallery input
   • Processes images with ML Kit
   • Filters results by configuration
   • Manages state lifecycle

3. Presentation Layer (ReusableResultCard)

   • Displays results with animations
   • Shows confidence levels visually
   • Provides responsive design

📖 API Reference

DetectionConfig

Parameter	Type	Default	Description
title	String	required	App title
instructions	String	'Take a photo'	User instructions
confidenceThreshold	double	0.5	Minimum confidence (0.0-1.0)
relevantLabels	List<String>	[]	Domain-specific filters
primaryColor	Color	Colors.blue	Theme color

ReusableMLDetector

Parameter	Type	Required	Description
config	DetectionConfig	Yes	Detection configuration

🛠️ Platform Setup

Android

Add to android/app/src/main/AndroidManifest.xml:

<uses-permission android:name="android.permission.CAMERA"/>
<uses-permission android:name="android.permission.READ_EXTERNAL_STORAGE"/>

<application>
    <meta-data
        android:name="com.google.mlkit.vision.DEPENDENCIES"
        android:value="image_labeling"/>
</application>

iOS

Add to ios/Runner/Info.plist:

<key>NSCameraUsageDescription</key>
<string>We need camera access to detect images</string>
<key>NSPhotoLibraryUsageDescription</key>
<string>We need photo library access to analyze images</string>

💡 Why This Component?

Not just a wrapper: This component demonstrates real reusability by abstracting complexity through external configuration. The same widget handles:

• ✅ Multiple detection domains without code duplication
• ✅ Consistent UI/UX across applications
• ✅ Centralized maintenance and updates
• ✅ Domain-specific filtering and validation
• ✅ Professional state management

Design Pattern: This approach is applicable to any ML or external API integration requiring multi-context adaptation.

🤝 Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

📄 License

This project is licensed under the MIT License - see the LICENSE file for details.

👨‍💻 Author

Dayver De la hoz

• GitHub: @DayverDelahoz
• University of Oviedo - UO312162

🙏 Acknowledgments

• Built with Google ML Kit
• Uses flutter_animate for animations
• Developed as part of Mobile Development coursework

📱 Example App

Check the /example folder for a complete demonstration app showing all four use cases.

🐛 Issues

Found a bug? Please open an issue.