CST: VoiceGrid

The multimodal system of criminalistic accounting and biometric search of VoiceGrid represents the product allowing to perform biometric search in two parameters – to the photographic image of the person and a voice.

The American analytical company Opus Research published in the fall of 2012 the report on a research of international market of voice biometrics in which the Center of Speech Technologies (CST) was recognized as world leaders in the field of multimodal biometrics.

Analysts highly appreciated the only Russian company presented in the analytical report, first of all, for a wide product range on sound processing and the speech of own development. The ability of the company to create the biometric systems of phonoaccounting and identification of the identity of national scale became one more important factor which displaid CST in leader positions. Such systems were created for law enforcement in the countries of Latin America.

For the first time the TsRT company created the biometric system of phonoaccounting and identification of the personality on a voice in 2010, thereby having created a new segment of the market of biometrics. Later two waiting in Ecuador one more similar system, the multimodal, integrated in itself sensing technologies of the personality on a voice and the photographic image of the person was created.

"The engine of voice biometrics developed in CST allows to make identification in large volumes, comparing millions of voice samples for a short period, - Dan Miller, the senior analyst in Opus Research says. – CST successfully copes with a task to qualitatively execute difficult, complex implementations of voice biometrics. Besides, it is the only company which managed to combine 2 types of biometrics in one solution".

The product VoiceGrid™ developed in CST serves for identification of the personality speaking, we scale and provides search in base to 10,000 records of voices. Solutions are the cornerstone algorithms of processing, segmentation and speech recognition and identification of the lines characteristic to a voice. In CST successfully integrated biometrics according to the photographic image of the person with these technologies, having received thereby a multimodal product.

The voice biometrics of CST also found the application and in the commercial sector, generally it is large contact centers. For this purpose in CST the line of solutions of VoiceKey including five different scenarios of application, everyone from which, it is directed to a certain type of problems of quality improvement of remote customer service and also increase in safety of the made transactions and safety of personal information was created. Thanks to new opportunities of confirmation of the personality on a voice in the automatic mode, contact centers received the convenient instrument of increase in efficiency of the work.

Advantages of a system

• Capture of biometric characteristics on the person and a voice is performed without contact, without participation investigated

• Unlike systems based on use of the only unique parameter of the person, a multimodal system allows to receive results of biometric search of high accuracy

• Low costs when collecting voice material and images of persons, deployment and operation of a system

• Both biometric technologies are ideally adapted for work on big databases

Data security provision

Reliable protection of information transferred and stored in a system is provided due to use of the https protocol, differentiation of user rights on the basis of groups and roles, registration of all actions of operators in base of audit and also protection of attendant workstation against theft of passwords.

Training

Specialists of our company perform complex personnel training of the Customer, providing the most high-quality and effective use of a system on places.

Technical support

The round-the-clock two-level technical support is performed by specialists at the central office of the developer and also local services on places.

Scalability

The lack of technical restrictions on volume of the stored information, by the number of users allows to use the solution as the universal platform for storage, search and other transactions with media and data within laboratory, the region or the whole state.

Possibility of remote work

The web interface, support of the majority of browsers in total with the protected data transfer protocols provide high degree of mobility of the specialist conducting investigation.

Platformonezavisimost

A system is compatible to all existing operating systems that allows to adapt to the Customer's IT infrastructure.

Integration with other criminalistic systems

Having the client-server architecture and user-friendly program interfaces, the system of biometric search is already ready to integration with other systems of both biometric, and not biometric accounting including expert systems of "The center of speech technologies" that allows to organize effective distributed systems of accounting, search and processing of any criminalistic data.

Work with material practically any quality

A system allows to carry out work with the voice and photographic material of practical any quality received from different sources.

2014: VoiceGrid X release 2.0.426

VoiceGrid X is the software product of CST for processing of telephone recordings. It is used for security of the enterprises and identification of telephone swindlers.

In release 2.0.426 the set of the improvements allowing to detect even more effectively in the automatic mode of a voice of target announcers (the swindlers issuing themselves for clients) in the written audiofiles thereby is implemented saving employees from routine work of 'manual' listening of all phonograms.

Significantly complex performance increased – now even on one core of the VoiceGrid X processor it is capable to process to 20,000 phonograms a day!

The last release included not only improvements of the existing functions, but also absolutely new opportunities:

• History module - the convenient tool is developed for filtering and viewing results of identification. Allows to filter results of work on a set of parameters and to listen to sound files with the found coincidence
• The reports generation subsystem is finished
• The ergonomics of the interface is improved