Face identification deception
Lack of control over spoofing-makes it easy to cheat the face identification system
The main problem of facial identification is to distinguish between twins
Solutions to detect the substitution of 2D face of a living person
The demo solution of detection of substitution of the live person on the Android smartphone (Chrome browser), on iPhone and Mac OS (Safari browser)
Solution demonstration
Demonstration program from Iris Devices for the Android operating system to detect the substitution of a living person
Download apk
Configuration instructions
The main application configuration options are summarized on a separate screen.
The main options are:
Protocol and address of access to the demo server.
In case of lack of access to the server-an error message will be displayed, and the result of identification will return the value: Fake.
The server address should be specified in the line: http://213:208:169:254:12001
This is the address of the demo server. The primary production server has a different address.
MinVerificationScore [decimal] – the minimum verification threshold [0..1].
MinIdentificationScore [decimal] – the minimum identification threshold [0..1].
UseAntispoofing [boolean] – use antispoofing (true).
MinSpoofingScore [decimal] – minimum threshold for detecting spoofing (face substitution) [0..1].
MinFaceSize [decimal] – the minimum face size to search in the image as a percentage of the maximum frame size.
MaxFaceSize [decimal] – the maximum face size to search in the image as a percentage of the maximum frame size.
IgnoreOtherFaces [boolean] – ignore the presence of several persons in the photo (true) or give an error of the presence of several persons in the frame (false).
MinFaceQuality [decimal] – the minimum quality threshold of the input face image to be added to the database [0..1].