Posted On: Dec 7, 2023
Amazon Rekognition Face Liveness enables application providers to detect in real time if a user accessing a service is a real person or a bad actor attempting to spoof faces to impersonate another identity or evade being recognized. Face spoofs can be presentation attacks (e.g., printed photos presented to a camera) or digital injection attacks (e.g., deep fake videos using software to bypass the camera). Amazon has launched a new version of Rekognition Face Liveness, which improves the model accuracy in detecting attacks and provides a new user interface (UI) to streamline the experience for genuine users.
Customers in various industries, such as financial, gig economy, telecommunications, healthcare, and social media, use Amazon Rekognition liveness detection and face matching for virtual onboarding, ongoing authentication, age-based access restriction, and bot detection. For example, application providers can conduct a liveness check for new user onboarding to confirm that it is a real person before further verifying their identity. The Rekognition Face Liveness update delivers higher accuracy to mitigate fraud and a simpler UI to reduce the liveness check time for genuine users.
To get the new user interface, application providers need to update their React, iOS, and Android SDKs. This update is available in all supported regions at no additional cost. To learn more about deploying Liveness responsibly and third party benchmark results from iBeta (ISO/IEC 30107-3), visit our Rekognition Face Liveness Service Card. To get started on implementation, please visit the Rekognition Face Liveness documentation.