Neurotechnology has made new product releases across their entire line of biometric Software Development Kits (SDK), including: Megamatcher 4.5 multi-biometric SDK for large-scale applications, Verifinger 6.7 for fingerprint identification, Verilook 5.4 for facial identification, Verieye 2.7 for iris identification and Verispeak 2.0 for voice identification.
The company has also added facial biometrics to the new Megamatcher Accelerator 6.0 multi-biometric all-in-one software and hardware solution. Now facial recognition can be used in any combination with fingerprint and iris biometrics for large-scale projects that require high-speed, high-volume identification capabilities.
The new facial biometric capabilities in Megamatcher Accelerator 6.0 offer fast matching speeds of up to 100 million templates per second and the solution has a capacity of up to 10 million faces on a single server.
Megamatcher Embedded SDKs are now included in Megamatcher SDK, which will enable users to develop PC, embedded or mixed biometric systems that support Windows, Linux, Mac OSX and Android platforms. Also, a new component, Megamatcher Accelerator Development Edition, has been added to the Megamatcher Extended SDK.
"Faces can now be matched at the same speed level as fingerprints and irises in our flagship product – Megamatcher Accelerator,” said Dr. Justas Kranauskas, project lead for Neurotechnology. “This will significantly reduce the complexity of large-scale identification/de-duplication projects that make use of facial biometrics. We are also hoping to lower the cost of development by adding a Development Edition of Megamatcher Accelerator to our Megamatcher SDK" Kranauskas continued.
With the merging of Megamatcher Embedded SDK into Megamatcher SDK, there will no longer be a separate Embedded SDK product. Now all customers who have obtained Megamatcher SDK will get cross-platform, embedded and mobile support for biometric systems on Windows, Linux, Mac OS X and Android without the need to purchase additional embedded SDKs.
The addition of Megamatcher Accelerator Development Edition to the Megamatcher Extended SDK allows Megamatcher SDK developers to create applications and solutions for Megamatcher Accelerator using the same functionality and programming interface (API) as the Megamatcher Accelerator Standard/Extended versions, although at lower matching speeds. Developing Accelerator-based products without need of a full version of Megamatcher Accelerator Standard or Extended will greatly reduce costs for Megamatcher Extended SDK customers.
Other changes in the product line include the Verispeak algorithm, which now allows for text-independent speaker recognition. In addition to the available passphrase recognition (text-dependent algorithm) in previous versions, Verispeak 2.0 includes a phrase-independent algorithm, which will allow for a wider range of application development. Verilook 5.4 has additional facial attribute recognition – smile detection, closed-eyes detection, open-mouth detection, glasses detection and dark-glasses detection,– that increases the potential application uses for the product.
The Megamatcher Accelerator 6.0 software/hardware solution is available in both Standard and Extended versions. Megamatcher Accelerator 6.0 Extended Version offers the highest performance for applications that require extremely fast matching with a large volume of data. The Extended version of Megamatcher Accelerator 6.0 is able manage up to 10 million faces and achieve a speed of 100 million faces per second. The fingerprint/AFIS and iris capabilities maintain the same features as Megamatcher Accelerator 5.0, with extremely fast matching speeds of up to 100 million fingerprints per second and up to 200 million irises per second. It can manage a database of up to 40 million fingerprints and 50 million irises. Multiple Megamatcher Accelerators can be connected in a cluster to manage databases with billions of records for even higher volume applications. Megamatcher Accelerator 6.0 Extended Version runs on an HP ProLiant DL360p G8 server with 128 GB RAM.
Megamatcher Accelerator 6.0 Standard Version is a fast, PC-based solution for customers with smaller biometric applications. Functionality is similar, however it manages approximately 1/10th the amount of data and is 1/3 as fast as the Extended version. Megamatcher Accelerator 6.0 Standard Version runs on a PC with an i7 processor, 16 GB RAM and 1.5TB HDD.
Face, fingerprint or iris modes can be used as a primary form of identification or any combination of the three modes can be used together as a multi-biometric solution. Megamatcher Accelerator maintains an extremely high level of accuracy because it performs a full database search for each match and does not use pre-classification or internal database indexing. As with the previous version, Megamatcher Accelerator 6.0 also provides a great deal of flexibility for the system developer to optimise for accuracy or speed and gives more control over the setting of engine parameters.
The ability to use multiple biometric modalities provides an even higher level of reliability than the use of any single biometric, and Megamatcher Accelerator provides the speed, accuracy and flexibility to scale to the millions of records required for national-level projects, such as: passport and visa documentation, citizen ID, voter registration, border control and criminal identification. It is also suitable for duplicate searches in large-scale government and commercial databases.