Flutter Face Detection

Detect with Firebase ML Kit + Draw with CustomPainter

In this article, we’ll explore the basics of detecting faces within an image with Firebase ML Kit and make it visible with the help of CustomPainter.

Workflow

• Get an image and convert it into a format that can be understood by our ML Kit.
• Feed the image to the detector and get it to scan the image for possible faces.
• Pull out the faces found and feed it to the CustomPainter.
• Allow the CustomPainter to get the coordinates of the faces and then use them to draw rects.

Dependencies:

We will need the image_picker and firebase_ml_vision packages to work out this project.

Implementation

Step 1 — Firebase Project and Dependencies

First create a Firebase project (refer to the docs) and add the image_picker and firebase_ml_vision dependencies to the pubspec.yaml file. Our dependencies should now look like this:

Also add the face model dependency to our app levelbuild.gradle file.

Step 2 — Fetch and Process the Image

Here, we first need to fetch the image from either the camera or the device gallery using the image_picker plugin.

final imageFile = await ImagePicker.pickImage(source:ImageSource.gallery);
seState((){
  isLoading = true;
});

Then let’s convert this image into bytes so it can be understood by the ML model and the custom painter as well.

final data = await file.readAsBytes();
await decodeImageFromList(data).then(
(value) => setState(() {
    _image = value;
    isLoading = false;
  }),
);

Step 3 — Detect the Faces

We will now create an instance of the Face Detector and task it to process the image to find any possible faces.

final faceDetector = FirebaseVision.instance.faceDetector();
List<Face> faces = await faceDetector.processImage(image);

Step 4 — CustomPainter

The face detector will return a list of Face class which contains the Rect coordinates. The Painter will then use these coordinates to draw a rectangular box around the faces using these coordinates. Below is the code for our CustomPainter.

Step 5 — Code the UI

We can now use the custom painter inside our widget tree like so:

Step 6— Enjoy

Now, let’s run the app and hope that we had everything right.

Yaaayyy!!! We’ve got our face detection app with some few lines of code thanks to the Flutter framework and Firebase.

Checkout out the source code of this project:

Akora-IngDKB/flutter_face_detection

Follow me on Twitter and feel free to send me a private message if you have a question:

Akora Ing. DKB

Flutter Community