MAD detects attempts to virtualize, intercept, or hook the device's camera, a practice in which the legitimate capture stream is replaced by images or videos from files, manipulated streams, or external sources.
This type of technique is widely used in frauds involving synthetic identity, deepfakes, and media manipulation, aiming to bypass authentication processes that rely on visual evidence, such as liveness checks, facial biometrics, and document validations in KYC (Know Your Customer) processes.
Technical Mechanism: MAD inspects the characteristics and behavior of the camera in use to verify whether the video stream originates from the device's physical sensor. The solution analyzes camera and capture environment properties, identifying inconsistencies common in virtualized cameras.
Upon identifying signs of manipulation or substitution of the video stream, MAD immediately applies the configured protection measures and reports the event in real time to the Command Center, ensuring the integrity of validation and authentication processes.
