RESOURCES AND CAPABILITIES:

• Tested on national scale projects, including passport issuance and deduplication of eleitor.
• Digital printing mechanism compatible with NIST MINEX, iris mechanism tested with NIST IREX.
• Multi-biometric turnkey solution for nationwide identification projects with MegaMatcher ABIS.
• High performance combination for large scale systems with MegaMatcher Accelerator.
• Digital prints, irises and faces can be combined on smart cards using MegaMatcher On Card.
• Includes digital printing, iris, face, voice and palm printing modalities.
• Laminated, flat and latent digital printing correspondence.
• BioAPI 2.0 and other ANSI and ISO biometric standards.
• Verification of compliance with the ICAO requirements for facial images.
• Effective price/performance relationship, flexible licensing and free customer support.

MEGAMATCHER DIGITAL PRINTING MODEL EXTRACTION AND CORRESPONDENCE MECHANISM.

Compliance with MINEX. NIST recognized the MegaMatcher digital printing algorithm as compatible with MINEX and suitable for use in personal identity verification program (PIV) applications.

Correspondence of laminated and flat digital prints. The MegaMatcher digital printing mechanism combines rolled and flat digital prints with each other. Normally, conventional “flat” digital print identification algorithms perform correspondence between flat and rolled digital prints in a less reliable manner due to specific deformations of rolled digital prints. MegaMatcher enables flat-to-flat, flat-to-laminated or laminated-to-laminated digital printing matching with a high degree of reliability and precision. The algorithm corresponds to 200,000 flat digital printing records per second on a single PC.

MegaMatcher includes a determination of the image quality of the digital print, which can be used during registration to ensure that only the best quality digital print model is stored in the database.

Detection of false digital impression. A classification of digitized digital print images based on deep learning is used to separate live/non-live digital prints to detect digital presentation attacks. This resource covers counterfeit attempts made with ecoflex, wood glue, latex and gelatin and is useful for identifying fraud.

The generalization of the model is used to generate a better quality model from several digital prints. Better quality models result in a higher level of identification accuracy.

MegaMatcher is tolerant of translation, rotation and deformation of digital printing. It uses a proprietary digital print matching algorithm that identifies digital prints even if they are rotated, moved or show deformations.

The adaptive image filtering algorithm eliminates noise, glass breaks and trapped grooves, and reliably extracts details at the same time as the lowest quality digital impressions in less than 1 second.

REMOVAL OF MEGAMATCHER FACIAL TEMPLATE AND CORRESPONDÊNCIA MECHANISM.

Model generalization is used to generate a better quality model from several face images. Better quality models result in a higher level of identification accuracy.

Tolerance in position guarantees a level of convenience in registration. MegaMatcher allows 360 degrees of head rotation. The head tilt can be up to 15 degrees in each direction starting from the front position. The head angle can be up to 45 degrees in each direction starting from the front position. See technical specifications for more details.

A reliable detection of face guarantor or precise registration of cameras, webcams and various digitized documents; Faces can be registered from digitized pages of passports or other types of documentation. When several faces are present in a video or image, they can be recorded and processed simultaneously. The sex of people, points of facial characteristics and basic emotions can be detected optionally. Also, partially occluded faces (such as people wearing masks or respirators) can be remade without separate registration.

Reconformation of facial attributes. MegaMatcher can be configured to detect certain attributes during face extraction – smile, open mouth, dark eyes, eyes, dark eyes, beard and mustache.

Estimate of ity. MegaMatcher can, optionally, estimate the identity of people by analyzing the face detected in the image.

Live face detection. A conventional face identification system can be fooled by placing a photo in front of the camera. MegaMatcher is capable of preventing this type of security breach by determining whether a face in a video stream is “alive” or in a photograph. Vivacity detection can be carried out in a passive mode, when the engine evaluates certain facial characteristics, and in an active mode, when the engine evaluates the user's response to perform actions such as clicking or head movements. See the recommendations for live face detection for more details.

The biometric model record may contain several facial samples belonging to the same person. These samples can be recorded from different sources and at different times, thus allowing for better quality of correspondence. For example, a person may be registered with eyeglasses or with different types of eyeglasses; with no beard or mustache, etc.

REMOVAL OF MEGAMATCHER VOICE TEMPLATE AND CORRESPONDÊNCI MECHANISM.

The text-dependent voice matching mechanism determines whether a voice sample corresponds to the model that was extracted from a specific phrase. During registration, one or more phrases are requested from the registered person. Later, this person may be asked to say a specific phrase for verification. This method guarantees protection against the use of a random phrase secretly recorded by people.

The authentication of two factors with a secret phrase is executed when a person is asked to say an exclusive phrase (such as a secret phrase or an answer to a “secret question” asked just by the person who is being registered). The general security of the system increases as the voice authenticity is verified.

The text-independent voice correspondence mechanism uses different phrases for user registration and recognition. This method is more convenient because it does not require each user to have a name. It can be combined with a text-dependent algorithm to perform a faster text-independent search with additional phrase verification using a more reliable text-dependent algorithm.

Automatic detection of voice activity. The mechanism is capable of detecting when users start and stop cheating.

Live detection. A system can request that each user enter a set of exclusive phrases. Subsequently, the user will be asked to say a specific phrase from the registered group. This way, the system can guarantee that a living person is being verified (as opposed to an imposter using voice recording).

Several voice recordings with the same phrase can be armed to improve the reliability of the loud-speaker recognition. Certain natural variations in voice (for example, rouca voice) or changes in environment (for example, desk and book) can be assembled in the same model.

EXTRAÇÃO MODEL OF ÍRIS MEGAMATCHER AND CORRESPONDÊNCI MECHANISM.

Proven reliability of NIST IREX. The MegaMatcher iris matching mechanism is based on VeriEye, remade by NIST as one of the two most reliable and accurate iris matching algorithms available.

Quick correspondence. The internet correspondence speed is around 200,000 comparisons per second on a single PC. See technical specifications for more details.

Robust iris detection. These are also detected when there are image obstructions, visual noise and/or different lighting levels. Illumination reflections, eyelid and cilia obstructions are eliminated. Images with dark eyelids and eyes that you are smelling for a long time also contain oil.

Automatic interlace detection and correlation results in maximum quality of iris resource models from moving images of irises.

Correct segmentation of the iris is obtained when the perfect circles are formed, the centers have two internal and external limits of the irises, the limits of the irises are definitely not circles and they are not ellipses or the limits of the irises look like perfect circles.

Determination of image quality and prevention of falsification. An estimate of image quality can be used during iris registration to ensure that only the best quality iris model is assembled in the database. Also, cosmetic (decorative) contact lenses, which obscure the iris with some artistic or false texture and/or alter the iris cord, can be detected.