Release notes for MegaMatcher 2025.x line products

These release notes are relevant for the following products:

• MegaMatcher 2025.1 Standard / Extended SDK
• VeriFinger 2025.1 Standard / Extended SDK
• VeriLook 2025.1 Standard / Extended SDK
• VeriEye 2025.1 Standard / Extended SDK
• VeriSpeak 2025.1 Standard / Extended SDK

You may subscribe for newsletters to receive notifications about the released updates.

See also the list of release notes for previous versions.

Product revision number: 20250304
Product revision hash: 03512a1628edab5ef0d55097320ba5f947f80f5b

The first release of MegaMatcher 2025.1 SDK line.

Algorithms-related changes:

• Fingers
  • New finger algorithms:
    • The new default finger algorithm (Fingerprint Matcher component license) is more accurate than the previous version.
    • An even better algorithm is available with the Fast Fingerprint Matcher component license.
    • A step further, MegaMatcher Accelerator allows achieving even better accuracy in complex cases.
    The algorithms or their prototypes have passed evaluations in PFT III (Neurotechnology_0013) and ELFT (Neurotechnology_0122).
  • A MINEX III compliant algorithm is available for ISO templates (listed as 010B submission).
  • New finger anomaly detector. Finger anomalies are various scanning issues, often related with user trying to cheat system and enroll multiple instances of biometrics (for example, footprints, different phalange, etc.).
  • Contactless fingerprints. These fingerprints are scanned from phone camera without using usual fingerprint scanners. New algorithms allow to transform such scans to usual fingerprint templates which can be later used for matching in compatible way.
• Faces
  • Improved faces recognition algorithm:
    • Better accuracy comparing to previous version (Face Extractor or Face Client component licenses).
    • Even more accurate Server extraction option available (works slower but produces better templates) with Face Image Processing component license,
    Algorithms or their prototypes were evaluated in FRTE 1-to-1 (Neurotechnology_015) and FRTE 1-to-N (Neurotechnology_014). For compatibility with the previous 12.x and 13.x versions, older face templates are automatically upgraded on the fly. In order to avoid repeating upgrade over and over again, the templates can be saved with upgraded features.
  • Significantly increased face tracking efficiency, accuracy, and functionality.
    • Newly expanded ICAO checks are now used by default for facial positioning, alignment, and best frame selection. This includes various new parameters that allow full control of face angle, distance, centerness, head movement, etc.
    • Automatic capture now supports six canonical positions as targets: legacy (using the old algorithm), portrait, half-profile (left and right), and full-profile (left and right). The system internally adjusts and provides target information based on the scenario by selecting one of the positions.
    • Added a new type of ICAO image crop for criminal and civil use, which accommodates various yaw and pitch angles, allowing for automatic face capture and alignment even in constrained environments. The thumbnail image was also updated to reuse ICAO cropping logic. It allows the creation of original, ICAO compliant, and small but high-quality thumbnail images for document printing, etc.
  • Face morph detection. Face morphs in biometrics are digitally altered images that blend facial features from two or more individuals. Neurotechnology developed morph detection algorithms, which analyze subtle inconsistencies in morphs, such as unnatural blending of features or pixel-level artifacts.
• Irises
  • More accurate iris recognition algorithm. In particular, more robust with non standard rotations and low quality data. Greatly improved performance when matching with high rotation angles. Algorithm prototype tested in IREX 10 (Neurotechnology_016).
• Palmprints
  • More accurate palmprint recognition algorithm. In particular, criminal complex cases gained most accuracy improvement.

New features:

• Added support of these ISO/IEC standards (Extensible biometric data interchange formats):
  • Finger image data (ISO/IEC 39794-4:2019);
  • Face image data (ISO/IEC 39794-5:2019);
  • Iris image data (ISO/IEC 39794-6:2021).

Updates and improvements:

• It is now possible to obtain licenses via HTTP/HTTPS (previously only through proprietary protocol). This allows user to configure any desirable proxies and setup desired level of additional encryption to meet organizational security requirements.