firebase_livestream_ml_vision

Description:

firebase livestream ml vision

ML Kit Vision for Firebase with AutoML Vision Edge and Camera Live Streaming Support #

A Flutter plugin to use the ML Kit Vision for Firebase API.
For Flutter plugins for other Firebase products, see FlutterFire.md.
Note: This plugin is still under development, and some APIs might not be available yet. Feedback and Pull Requests are most welcome!
Usage #
To use this plugin, add firebase_livestream_ml_vision as a dependency in your pubspec.yaml file. You must also configure Firebase for each platform project: Android and iOS (see the example folder or https://codelabs.developers.google.com/codelabs/flutter-firebase/#4 for step by step details).
AutoML Vision Edge #
If you plan to use AutoML Vision Edge to detect labels using a custom model, either download or host the trained model by following these instructions.
If you downloaded the file, follow the instructions below to enable the plugin.
Unzip the file downloaded, and rename the folder to reflect the model name.
Create a assets folder and place the previous folder within it. In pubspec.yaml add the appropriate paths:
assets:
- assets/<foldername>/dict.txt
- assets/<foldername>/manifest.json
- assets/<foldername>/model.tflite
copied to clipboard
Android #
Change the minimum Android sdk version to 21 (or higher) in your android/app/build.gradle file.
minSdkVersion 21
copied to clipboard
If you're using local AutoML VisionEdge Models, include this in your app-level build.gradle file.
aaptOptions {
noCompress "tflite"
}
copied to clipboard
If you're using the on-device ImageLabeler, include the latest matching ML Kit: Image Labeling dependency in your app-level build.gradle file.
android {
dependencies {
// ...

api 'com.google.firebase:firebase-ml-vision-image-label-model:17.0.2'
}
}
copied to clipboard
If you receive compilation errors, try an earlier version of ML Kit: Image Labeling.
Optional but recommended: If you use the on-device API, configure your app to automatically download the ML model to the device after your app is installed from the Play Store. To do so, add the following declaration to your app's AndroidManifest.xml file:
<application ...>
...
<meta-data
android:name="com.google.firebase.ml.vision.DEPENDENCIES"
android:value="ocr" />
<!-- To use multiple models: android:value="ocr,label,barcode,face" -->
</application>
copied to clipboard
iOS #
Versions 0.7.0+ use the latest ML Kit for Firebase version which requires a minimum deployment
target of 9.0. You can add the line platform :ios, '9.0' in your iOS project Podfile.
If you're using one of the on-device APIs, include the corresponding ML Kit library model in your
Podfile. Then run pod update in a terminal within the same directory as your Podfile.
pod 'Firebase/MLVisionBarcodeModel'
pod 'Firebase/MLVisionFaceModel'
pod 'Firebase/MLVisionLabelModel'
pod 'Firebase/MLVisionTextModel'
copied to clipboard
Add one row in ios/Runner/Info.plist:
The key Privacy - Camera Usage Description and a usage description.
In text format:
<key>NSCameraUsageDescription</key>
<string>Can I use the camera please?</string>

copied to clipboard
Using an ML Vision Detector #
1. Create a Camera View #

import 'package:firebase_livestream_ml_vision/firebase_livestream_ml_vision.dart';
import 'package:flutter/material.dart';

void main() => runApp(MaterialApp(home: _MyHomePage()));

class _MyHomePage extends StatefulWidget {
@override
_MyHomePageState createState() => _MyHomePageState();
}

class _MyHomePageState extends State<_MyHomePage> {
FirebaseVision _vision;
dynamic _scanResults;

@override
void initState() {
super.initState();
_initializeCamera();
}

void _initializeCamera() async {
List<FirebaseCameraDescription> cameras = await camerasAvailable();
_vision = FirebaseVision(cameras[0], ResolutionSetting.high);
_vision.initialize().then((_) {
if (!mounted) {
return;
}
setState(() {});
});
}
Widget _buildImage() {
return Container(
constraints: const BoxConstraints.expand(),
child: _vision == null
Stack(
fit: StackFit.expand,
children: <Widget>[
FirebaseCameraPreview(_vision),
_buildResults(),
],
),
);
}

Widget build(BuildContext context) {
return Scaffold(
body: _buildImage(),
);
}

void dispose() {
_vision.dispose().then((_) {
// close all detectors
});

super.dispose();
}
}

copied to clipboard
See the example app to learn more on incorporating detectors in the camera app, check it out here.
2. Using detectors #
Special Instructions for using VisionEdgeImageLabeler #
Get an object of ModelManager, and setup the local or remote model(optional, results in faster first-use)
FirebaseVision.modelManager().setupModel('<foldername(modelname)>', modelLocation);
copied to clipboard
Calling a Labeler/Detector
a. Image Labeler
_vision.addImageLabeler().then((onValue){
onValue.listen((onData) => // do something with data
);
});
copied to clipboard
b. Cloud Image Labeler
_vision.addCloudImageLabeler().then((onValue){
onValue.listen((onData) => // do something with data
);
});
copied to clipboard
c. Barcode Detector
_vision.addBarcodeDetector().then((onValue){
onValue.listen((onData) => // do something with data
);
});
copied to clipboard
d. Face Detector
_vision.addFaceDetector().then((onValue){
onValue.listen((onData) => // do something with data
);
});
copied to clipboard
e. Text Recognizer
_vision.addTextRecognizer().then((onValue){
onValue.listen((onData) => // do something with data
);
});
copied to clipboard
f. Vision Edge Image Labeler
_vision.addVisionEdgeImageLabeler('<foldername(modelname)>', modelLocation).then((onValue){
onValue.listen((onData) => // do something with data
);
});
copied to clipboard
You can also configure all detectors, except TextRecognizer, with desired options.
final ImageLabeler labeler = FirebaseVision.instance.addImageLabler(
ImageLabelerOptions(confidenceThreshold: 0.75),
);
copied to clipboard
3. Extract data #
a. Extract barcodes.
for (Barcode barcode in barcodes) {
final Rectangle<int> boundingBox = barcode.boundingBox;
final List<Point<int>> cornerPoints = barcode.cornerPoints;

final String rawValue = barcode.rawValue;

final BarcodeValueType valueType = barcode.valueType;

// See API reference for complete list of supported types
switch (valueType) {
case BarcodeValueType.wifi:
final String ssid = barcode.wifi.ssid;
final String password = barcode.wifi.password;
final BarcodeWiFiEncryptionType type = barcode.wifi.encryptionType;
break;
case BarcodeValueType.url:
final String title = barcode.url.title;
final String url = barcode.url.url;
break;
}
}
copied to clipboard
b. Extract faces.
for (Face face in faces) {
final Rectangle<int> boundingBox = face.boundingBox;

final double rotY = face.headEulerAngleY; // Head is rotated to the right rotY degrees
final double rotZ = face.headEulerAngleZ; // Head is tilted sideways rotZ degrees

// If landmark detection was enabled with FaceDetectorOptions (mouth, ears,
// eyes, cheeks, and nose available):
final FaceLandmark leftEar = face.getLandmark(FaceLandmarkType.leftEar);
if (leftEar != null) {
final Point<double> leftEarPos = leftEar.position;
}

// If classification was enabled with FaceDetectorOptions:
if (face.smilingProbability != null) {
final double smileProb = face.smilingProbability;
}

// If face tracking was enabled with FaceDetectorOptions:
if (face.trackingId != null) {
final int id = face.trackingId;
}
}
copied to clipboard
c. Extract labels.
for (ImageLabel label in labels) {
final String text = label.text;
final String entityId = label.entityId;
final double confidence = label.confidence;
}
copied to clipboard
c. Extract Cloud Vision Edge labels.
for (VisionEdgeImageLabel label in labels) {
final String text = label.text;
final double confidence = label.confidence;
}
copied to clipboard
d. Extract text.
String text = visionText.text;
for (TextBlock block in visionText.blocks) {
final Rect boundingBox = block.boundingBox;
final List<Offset> cornerPoints = block.cornerPoints;
final String text = block.text;
final List<RecognizedLanguage> languages = block.recognizedLanguages;

for (TextLine line in block.lines) {
// Same getters as TextBlock
for (TextElement element in line.elements) {
// Same getters as TextBlock
}
}
}
copied to clipboard
Getting Started #
See the example directory for a complete sample app using ML Kit Vision for Firebase with Camera Streaming.