Speed and reliability…
The purpose of this SDK is to provide functionality to manage cameras (both video and static) and perform face detection. It also provides feature extraction (quality control) and subsequent facial recognition: both in 1:1 and 1:N engines.
The SDK is designed for the simple integration of facial recognition functionality. It can be used both in security applications that require biometric identity verification, as in other types of solutions. The SDK can act autonomously or access a high-level API that offers interconnection with other systems.
The process starts with a camera connected to the computer. Frames and images are extracted for processing: facial recognition, coincidence and quality control.
Quality – Processing
It is the phase in which a subject can be detected by evaluating an image, where the person is present. The face detection process recovers the location of the face with a high degree of accuracy. Even in cases where he has a beard, is turned, or wear glasses.
Face QA: It allows to evaluate if an image is compatible with the ICAO requirements verifications. The different verifications carried out:
Verticality: the face is vertically aligned with the image.
Horizontality: the face must be centered horizontally.
Good exposure: the face must be subject to good lighting to be detected correctly.
Red eyes: for a good facial recognition it is necessary that the eyes do not look red, but of their natural color.
Resolution: the resolution of the image must allow the application to clearly detect the subject in question.
Only one visible face: It is necessary to detect a single face, to identify and analyze successfully the individual in question.
Face matching is a phase in which an individual can be identified by evaluating their biometric characteristics. When comparing with another image, it is verified if the subject is the person he claims to be. This process is called one-to-one correspondence because both images are of the same individual. There is also the possibility of comparing images with several, which is what we call pairing from one to many.
1:1 Verification of the match between a document and a person.
1 to N, identification and search of a person to N.
Find out what can we do for you
In NeoCheck® we strive to satisfy all the needs of our customers in terms of Document Verification and Biometric Identification. From web-based solutions, mobile applications to specialized components (we have our own research and development team). And of course, we try to deploy the most advanced and flexible technology at affordable prices, as well as providing them with the best support. Therefore, we regularly organize online courses and workshops related to the world we know best: Document and ID Verification.