🔥 Checkout our main example to get the right way to start.

Get Started

~~~
npm install react-native-facerecognition --save
or
yarn add react-native-facerecognition
~~~
>⚠️ We are highly recommend you to use Deamon gradle for building faster and ignore any error may happen.
* Inside build.gradle put this line in projectName/android/build.gradle.
~~~
repositories {
...
google()
}
...
classpath 'com.android.tools.build:gradle:3.0.1'
...
allprojects {
repositories {
...
maven { url "https://jitpack.io" }
maven { url "https://maven.google.com" }
...
}
}

~~~
* And go to projectName/android/gradle/wrapper/gradle-wrapper.properties and change this.
~~~
distributionUrl=https\://services.gradle.org/distributions/gradle-4.4-all.zip
~~~

$3

* You would need to link the library once you install it.
~~~
react-native link react-native-facerecognition
~~~
* Add this line into build.gradle and the path projectName/android/app/build.gradle.
~~~
compileSdkVersion 26
buildToolsVersion "26.0.2"
....
....
~~~
* Then put this lines into settings.gradle and the path projectName/android/settings.gradle.
~~~
include ':openCV'
project(':openCV').projectDir = new File(rootProject.projectDir, '../node_modules/react-native-facerecognition/android/openCV')
~~~
* You need to put permissions to use Camera on AndroidManifest.xml and remove the line below.
~~~





...
android:windowSoftInputMode="adjustResize" // remove this in order to make UI works well
~~~

$3

* First, put this line into MainApplication.java and its path projectName/android/app/src/main/java/com/projectName/MainApplication.java.
~~~
import opencv.android.FaceModulePackage;
...
return Arrays.asList(
new FaceModulePackage()
);
~~~
* Then add this line into build.gradle and the path projectName/android/app/build.gradle.
~~~
compileSdkVersion 26
buildToolsVersion "26.0.2"
....
....
....
dependencies {
compile project(':react-native-facerecognition')
}
~~~
* Inside settings.gradle put these lines you can find the path at projectName/android/settings.gradle
~~~
include ':react-native-facerecognition'
project(':react-native-facerecognition').projectDir = new File(rootProject.projectDir, '../node_modules/react-native-facerecognition/android/app')
include ':openCV'
project(':openCV').projectDir = new File(rootProject.projectDir, '../node_modules/react-native-facerecognition/android/openCV')
~~~
* You need to put permissions to use Camera on AndroidManifest.xml and remove the line below.
~~~





...
android:windowSoftInputMode="adjustResize" // remove this in order to make UI works well
~~~

🔥 You're ready to go!

Documentation

$3

model
* Camera.constants.Model.cascade [DEPRECATED]
* Higher recall & More trained faces
* Camera.constants.Model.landmarks
* Higher accuracy at recognition
* Camera.constants.Model.lbp
* Higher precision & Faster & Less trained faces

cameraType
* Camera.constants.CameraType.front
* Camera.constants.CameraType.back

captureQuality
* Camera.constants.CaptureQuality.low
* Camera.constants.CaptureQuality.medium
* Camera.constants.CaptureQuality.high

aspect
* Camera.constants.Aspect.fit
* Camera.constants.Aspect.fill
* Camera.constants.Aspect.stretch

dataset (default: false)
* train your own images. copy Images to that location at android/app/src/main/assets/dataset inside that folder.
> if you will use this, you need to use lbp model only in order to make it work. Please see the test photo there to know how to rename your photos according to that test photo every photo should be named like that name_number.jpg/png.

distance
~~~

~~~
* Distance between the face and the camera. This is very important to keep recognition always works and to help make the result mainly true (this effects the recognition result grossly).

rotateMode - (Landscape/Portrait)
* Camera.constants.RotateMode.on
* Camera.constants.RotateMode.off

$3

* takePicture()
> Take a picture then process it to detect face inside

* train(Object)
> Train the algorithm with a new detected face

* identify()
> Take a picture then predict whom face belongs to

* clear()
> Clear all previous trained faces

$3

onTrained
* Called after success training

onUntrained
* If training fails this function will be called with the error

onRecognized
You get details about recognized face:
* name The face name
* confidence This number indicates how much the result is true. Usually low number < 100 means a good result depends on the distance you put

onUnrecognized
* If recognition fails this function will be called with the error

Recommendations

>If you're not familiar with OpenCV and face recognition you have to be in safe and use our default arguments as we care about all of the details for you. In case of using your own arguments please note that you may effect the accuracy depends on your settings.

* Recognition default arguments

__Arguments__
- distance = {200}

* Training images count

__Minimum__
- There's no maximum but the average is 3~4 photos per face to guarantee high accuracy

Updates

🚀 Updates you can find all previous versions and updates up to date!

Thank to

rn-openALPR
rn-camera

Licence

MIT