Netraksh AI

Offline Facial Recognition & Liveness Detection System for Remote Locations

Netraksh AI is an offline-first facial authentication and liveness workflow built with React Native 0.72.7, TypeScript, and SQLite. It is designed for remote environments where connectivity is unreliable, keeping enrollment, face matching, liveness challenge state, and local audit logging available on device.

---

Features

• Camera Capture: Uses react-native-vision-camera for enrollment capture, authentication capture, permission handling, retake flow, and development mock capture.
• Face Embeddings: Uses ArcFace-MobileNetV2-style 512-dimensional normalized embeddings with react-native-fast-tflite integration and deterministic mock fallback when native inference is unavailable.
• Liveness Challenge Flow: Implements blink, smile, and head-turn challenge state logic. The current UI feeds simulated EAR/MAR/yaw metrics; real landmark detection is still a production hardening item.
• Encrypted Face Templates: Stores encrypted embeddings in SQLite using AES-256-CBC and a Keychain-backed local key. Current encryption still needs authenticated encryption and secure-random cleanup before production use.
• Offline Audit Logs: Stores authentication attempts locally with sync status, device/location context, and SHA-256 integrity hashes for tamper evidence.
• Offline-to-Online Sync: Provides manual sync from the Offline Logs screen, automatic sync on network restore, and periodic scheduled sync using AUTH_LOG_SYNC_INTERVAL_MS.
• AWS Backend Scaffold: Includes a SAM/API Gateway/Lambda/DynamoDB sync backend with request validation. Backend authentication and deployment validation are still required.

---

Setup & Installation

Prerequisites

• Node.js: v18.0.0 or higher
• Java SDK: JDK 21
• Android: Android Studio with Android SDK and command-line tools
• iOS: macOS with Xcode and CocoaPods installed

Installation

1. Clone the repository:

   git clone https://github.com/anjali9in/Netraksh-AI.git
   cd Netraksh-AI

2. Install dependencies:

   npm install

3. Install Pods (iOS only):

   npm run ios:pods

---

Running the App

Running on Android

Ensure an emulator is running or a physical device is connected via USB with USB debugging enabled.

# Start Metro bundler
npm start

# Run on Android
npm run android

Running on iOS

# Run on iOS Simulator
npm run ios

---

Developer Commands

# Run Jest Unit Tests
npm test

# Run TypeScript Type-Checker
npm run typecheck

# Run ESLint Code Linter
npm run lint

---

📁 Project Architecture

• src/app: Navigation routes, navigator screens, and shell framework.
• src/screens: Views for enrollment, authentication, logs history, benchmarks, and settings.
• src/components: Reusable UI elements (CameraCaptureCard, LiveScannerPanel, Navigation bars).
• src/services: Database repositories, liveness state machines, geolocation, and network services.
• src/ai: TFLite generator, preprocessing, similarity metrics, and benchmark scripts.
• src/utils: Encryption helper, hash validation, and date functions.
• backend: AWS SAM sync API scaffold.
• docs: Architecture, setup, benchmark, testing, and next-step notes.
• __tests__: Jest tests covering validation, routes, device info, camera orientation, liveness, offline database behavior, and AI utilities.