digit_face_verification 0.0.1

Offline face verification for field worker attendance using MobileFaceNet + Google ML Kit.

digit_face_verification

Offline face verification for field worker attendance using MobileFaceNet + Google ML Kit.

The package provides:

• On-device face enrollment and verification (no server round-trip required)
• Liveness detection to defeat photo/video replay
• Distance-metric based match scoring against a locally-stored embedding
• A ready-to-use capture, enrollment, and verification UI

Getting Started

Add the dependency to your app's pubspec.yaml:

dependencies:
  digit_face_verification: ^0.0.1

Register the BLoCs your flow needs — typically enrollment, verification, and liveness — near the root of your widget tree:

MultiBlocProvider(
  providers: [
    BlocProvider(
      create: (_) => FaceEnrollmentBloc(),
    ),
    BlocProvider(
      create: (_) => FaceVerificationBloc(),
    ),
    BlocProvider(
      create: (_) => LivenessBloc(),
    ),
  ],
  child: const MyApp(),
);

Enrolling a user

context.router.push(
  FaceEnrollmentRoute(individualId: individual.clientReferenceId),
);

Verifying a user

context.router.push(
  FaceVerificationRoute(individualId: individual.clientReferenceId),
);

Wrapping attendance actions

For attendance flows, wrap the guarded action in FaceAttendanceWrapper so the user must pass face verification before the action fires:

FaceAttendanceWrapper(
  individualId: individual.clientReferenceId,
  onVerified: () {
    // Mark attendance / trigger the guarded action
  },
  child: const AttendanceMarkButton(),
);

Configuration

• The MobileFaceNet model and audio prompts ship as bundled assets under assets/models/ and assets/audio/.
• Match-cutoff, liveness challenge sequence, and localization keys are driven from the app's MDMS face-auth config.