Introduction
v1.4.0
Last updated
v1.4.0
Last updated
Welcome to our documentation of our state-of-the-art Passive Facial Liveness Detection REST API. The SpoofSense Face Liveness API system offers advanced computer vision and AI-based passive facial liveness detection to help you detect face spoofing attacks and authenticate live users using just a selfie. Our Passive Liveness Detection algorithm is tested by iBeta and is compliant with Level-2 ISO 30107-3 standards. Read our conformance letter .
Improved Accuracy: Enhanced model performance ensures even more precise liveness detection.
Reduced False Negative Rate: Significant reduction in false negatives improves reliability in identifying genuine users.
Increased Robustness: Better handling of varied scenarios, ensuring consistent and accurate results.
Spoofs or Presentation Attacks are representations of human faces used by fraudsters to fool face biometric systems and carry out identity theft. SpoofSense Facial Liveness distinguishes a live person from spoofing attacks. This process is called Liveness Detection or Presentation Attack Detection (“PAD”). Unlike most other systems that perform PAD, SpoofSense Facial Liveness does not require any action or movement. It requires just a single-frame selfie image, the same image that is often used by facial biometric systems to carry out face matching. SpoofSense Facial Liveness is a passive liveness solution. Fraudsters use the following Presentation Attacks for spoofing:
Printed Photo Attack: A fraudster prints photos of an individual and presents them in front of the camera during verification.
Video Replay Attack: A fraudster uses a digital screen of a phone or an iPad and presents a video of an individual being played on the screen in front of the camera to fool the verification system.
Printed Mask Attack: A fraudster places a cut-out photo in front of their face during verification, often with cut-out holes so the impostor can blink, a common test of liveness for other systems.
3D Mask Attack: A fraudster creates a silicon or resin mask of an individual and presents it in front of their face during verification.
SpoofSense Facial Liveness requires just a single image of the user, taken from the front camera of their phone. SpoofSense API returns a score and probability to help you make decisions about liveness.
Passive Liveness Detection: No user interaction or movements are required; a single selfie suffices.
High Security Standards: Tested and compliant with ISO 30107-3 Level-2 standards.
Wide Range of Attack Detection: Robust against various spoofing techniques such as photo, video, and mask-based attacks.
The remainder of this documentation describes technical requirements, image sample guidelines, and instructions on how to integrate and run the SpoofSense Facial Liveness API.
