Release notes for MegaMatcher Accelerator 13

See also the list of release notes for previous versions.

Algorithms-related changes:

• Improved accuracy for latent palm.

New features:

• Get subject count per gallery.

Fixes:

• Fixed several small bugs and issues.

Updates and improvements:

• Added metrics for database sizes and capacities.
• Added metrics for database loading speed diagnostics.
• Upgraded third party dependencies to avoid possible security vulnerabilites.
• Support for externals databases which can occasionally disconnect. With Retriable failover configuration, MMA node waits for database to come back instead of restarting.
• Statistics loading in WebConfig a lot faster.

The first release of MegaMatcher Accelerator 13.1.

Algorithms-related changes:

• New voice matching algorithm is now available for accelerated matching in MegaMatcher Accelerator product. This opens new opportunities for processing identification in large database in very quick time. Also, voice modality has been integrated into MegaMatcher ABIS product.
• Algorithms updated with 13.1 product line improvements.

Fixes:

• Fixed an issue with large database loading.

Fixes:

• Fixed WebConfig issues.
• Fixed several small bugs and issues.

Updates and improvements:

• Added support for secure encrypted gRPC communication.
• More optimization parameters for multimodal scenarios.

Fixes:

• Fix: truncate too large matching results.
• Fixed several small bugs and issues.

Fixes:

• Fixed issues with accelerated voices matching speed.
• Fixed GetCapacityInfo command.

Updates and improvements:

• Aligned MariaDB version to be 10.6 – the same as in MegaMatcher ABIS product.
• Improved WebConfig user experience for voice accelerator information and statistics.

Fixes:

• Fixed an issue with iris engine mixing old and new templates during insert/delete operations, which may result in incorrect operation.

The first release of MegaMatcher Accelerator 13.0.

Algorithms-related changes:

• Major upgrades to fingerprint, face, iris, and palmprint algorithms for fast and highly accurate identification using single or multiple biometric modalities.
• Voices modality can also be accelerated since version 13.0 which opens possibility for much faster identification speed than previously.