Neurotechnology’s new MegaMatcher performs 1.2B biometric matches in one second without GPU
The latest version of Neurotechnology’s multi-biometric MegaMatcher portfolio of solutions and SDKs has been released with what the company calls “major upgrades” to its fingerprint, face, iris and voice recognition algorithms for fast and accurate identification.
In addition to algorithm enhancements, the new MegaMatcher 12 includes an enhanced MegaMatcher Accelerator. The Extreme edition can match up to 1.2 billion fingerprints, faces or irises per second without requiring a GPU. This capability simplifies deployment, the company says, and does not involve a biometric accuracy trade-off.
Each modality can be used in MegaMatcher 12 alone or in combination for fast and reliable identification.
“The MegaMatcher 12 line reflects the results of several years of achievements in our research and development, with improvements that make it easier to use biometrics at a large scale,” says Irmantas Naujikas, director for Neurotechnology. “In addition to releasing this upgrade, we are also pleased to highlight that our algorithms continue to achieve top rankings by accuracy in the latest independent evaluations – including the first place in both NIST MINEX III and NIST PFT III, the second place in NIST IREX IX, the first place in FVC-onGoing fingerprint and palmprint verification – and, most importantly, they are proven in national scale biometric projects worldwide.”
Other notable enhancements in the new release include an improved ability to detect incorrectly scanned images and finger liveness detection for fraud prevention. Face detection and feature points detection has been upgraded, iris matching speed has been increased, and the latest voice algorithm performs improved test-independent recognition.
The MegaMatcher line includes MegaMatcher ABIS, Accelerator and SDK.
Neurotechnology recently released a comprehensive hardware and software development kit for EEG biometrics along with an update to its SentiSight.ai object recognition platform.