Liveness Detection

Autohost’s Liveness Detection is an advanced component of our ID Verification solution that provides an essential layer of security against presentation attacks and identity fraud.

Unlike basic selfie verification that can be fooled with printed photos or screen displays, Liveness Detection uses sophisticated computer vision technology to confirm that a real person is physically present during the verification process.

After confirming liveness, the system securely sends facial images to our API for comparison against the user’s ID photo and optionally against Global Watchlist databases, providing a comprehensive identity verification solution that protects your business from fraud and unauthorized access.

Our Liveness Detection process combines advanced technology with a simple user experience:

1. Video Stream Initialization: The system activates the user’s device camera and starts a secure video stream.

2. On-Device Liveness Analysis: Using WebAssembly (WASM) technology, a lightweight AI model runs on the user’s device to track facial movements and verify physical presence.

3. Movement Guidance: Users are prompted to perform simple head rotations from side to side, following on-screen instructions.

4. Real-Time Analysis: As the user moves, the system analyzes facial movements, depth changes, and three-dimensional characteristics that can only be produced by a live person.

5. Secure Image Transmission: Once liveness is confirmed, facial images are securely sent to Autohost’s API for further verification.

6. ID Comparison: The API compares the captured facial images against the photo on the user’s ID document to confirm identity.

7. Optional Watchlist Check: If enabled, the system also compares facial images against Global Watchlist databases for additional security.

Liveness Detection provides multi-layered identity verification:

• Presentation Attack Prevention: Confirms the user is physically present and not using photos, videos, or masks to spoof the system.

• Biometric Identity Matching: Compares facial features with ID document photos to verify the user is the legitimate ID holder.

• Global Watchlist Screening: Optionally checks facial images against international watchlists to identify potential high-risk individuals.

• Comprehensive Risk Assessment: Integrates verification results with Autohost’s Screening AI for holistic risk evaluation.

This multi-faceted approach ensures that users are who they claim to be, are physically present during verification, and do not pose known security risks—all within a seamless verification flow.

Liveness Detection protects against sophisticated presentation attacks:

• Photo Spoofing: Detects when someone attempts to use a printed or digital photo instead of their actual face.

• Screen Replay Attacks: Identifies when pre-recorded videos are displayed on screens to try to fool the verification system.

• Deepfake Prevention: Analyzes micro-movements and natural facial patterns that are difficult for even sophisticated deepfakes to replicate.

• Identity Theft Protection: By comparing faces against ID photos, the system prevents individuals from using stolen or borrowed identification documents.

• High-Risk Individual Screening: The optional watchlist comparison helps identify individuals who may pose security risks.

Autohost’s Liveness Detection is built with security and privacy as core principles:

• Hybrid Processing Approach: Liveness detection happens on-device for privacy, while secure facial comparison occurs via our encrypted API for enhanced security.

• Encrypted Transmission: All facial images sent to our API are transmitted using industry-standard encryption protocols.

• Secure Storage: Any stored facial data is protected with enterprise-grade security measures and strict access controls.

• Transparent Process: Users are clearly informed about what data is collected, how it’s used, and who it’s shared with.

• Compliance-Ready: Our systems are designed to comply with privacy regulations including GDPR, CCPA, and biometric information laws.

• Data Minimization: We collect only the data necessary for verification purposes and maintain appropriate retention policies.