KYC Aura Scan

AI-Powered Video KYC with Multi-Modal Deepfake Detection

Problem Context

Digital identity verification is under threat from AI-generated deepfakes and replay attacks. Existing Video KYC solutions often rely on single-layer liveness checks (e.g., blink detection), which can be bypassed. Hybrid ecosystems (cloud + on-prem) require continuous identity assurance that is both secure and scalable.

Core Innovation

Our solution introduces a multi-modal, parallelized deepfake-resistant verification pipeline that combines:

• Behavioral cues (prompted liveness actions, micro-dynamics)
• Visual forensics (GAN artifact detection, 3D consistency checks)
• Temporal integrity (loop/replay detection)
• Cross-modal validation (lip–audio sync for prompted phrases)
• ID binding (face-to-ID photo verification)

The innovation lies in running all checks independently and in parallel, then fusing results into a probability matrix that feeds a fusion engine for final decision-making.

Key Features

Parallelized Multi-Check Engine

Ensures no single-point failure; deepfakes fail across multiple dimensions.

Active + Passive Defenses

• Active: random prompts (blink, head turn, read digits)
• Passive: forensic CNN spotting GAN fingerprints

Cross-Modal Fusion

Lip movements aligned with audio speech to defeat dubbed/fake audio overlays.

Temporal Security Layer

Replay detection using frame hashing and fps anomaly detection.

Realtime Probability Matrix Dashboard

Judges and auditors can see in real time how each check scores.

Signed Audit Trail

Each session produces a JSON report with SHA256 integrity hash for compliance.

Getting Started

Prerequisites

• Node.js 18+
• npm or yarn

Installation

1. Clone the repository:

git clone <repository-url>
cd kyc-aura-scan

2. Install dependencies:

npm install

3. Start the development server:

npm run dev

The application will be available at http://localhost:8080

Available Scripts

• npm run dev - Start development server
• npm run build - Build for production
• npm run build:dev - Build for development
• npm run lint - Run ESLint
• npm run preview - Preview production build

Technology Stack

• Frontend: React 18 + TypeScript + Vite
• UI Components: shadcn/ui + Tailwind CSS
• State Management: TanStack Query
• Form Handling: React Hook Form + Zod
• Charts: Recharts
• Icons: Lucide React

Project Structure

src/
├── components/         # Reusable UI components
│   ├── ui/            # shadcn/ui components
│   ├── ControlPanel.tsx
│   ├── PromptCard.tsx
│   └── ScoreChart.tsx
├── pages/             # Route components
│   ├── Home.tsx       # Landing page
│   ├── Upload.tsx     # Video upload
│   ├── Live.tsx       # Live camera KYC
│   └── Dashboard.tsx  # Analysis dashboard
├── hooks/             # Custom React hooks
├── lib/               # Utility libraries
└── utils/             # Business logic
    └── mockEngine.ts  # KYC simulation engine

Architecture

The application simulates a comprehensive Video KYC system with:

• Multi-modal verification checks running in parallel
• Real-time scoring dashboard with live probability updates
• Session management with unique IDs and audit trails
• Responsive design optimized for both desktop and mobile
• Mock simulation engine demonstrating production-level KYC workflows

License

This project is licensed under the MIT License.