How dqh147258/facerecognition works
Project type and purpose
• This is a mobile face-recognition authentication sample for Android, composed of two Gradle modules:
• “app” – a reference Android application (Activities, layouts, camera UI).
• “facerecognition” – a Kotlin/Android library implementing the ML face-recognition engine (uses TensorFlow Lite).
• The README confirms the goal: “Face recognition for authentication (1:1)”.
File structure
• Top-level modules:
• app/
– UI Activities under src/main/java (MainActivity, AddFaceActivity, FaceRecognitionActivity)
– Android resources (layouts, drawables, themes)
– Tests under src/androidTest and src/test
• facerecognition/
– Core classes: FaceRecognition.kt, FaceModelManager.kt
– Processor subpackage (chain of pipeline handlers for detection, alignment, comparison, etc.)
– tflite subpackage (TensorFlowLiteAnalyzer)
– model & util subpackages
– assets (mobile_face_net.tflite)
– Tests under src/androidTest and src/test
• Build scripts: build.gradle in each module, settings.gradle, wrapper under gradle/, proguard rules.
README
• Confirms 1:1 face-recognition focus, no server side. All processing happens on device with the .tflite model.