Human Speech Recognition Technology

ASR (Automatic Speech Recognition)

It is a translation technology of the sound waves represented by a human speech to digital information in the form of text symbols (plain text). Application and use:

• Voice control of different appliances and equipment
• Voice assistants to be integrated within telephone systems and mobile devices
• Text production using human speech recognition technology

NLU (Natural Language Understanding)

It is a human speech recognition technology. It may be used for the following applications:

• Identification of meaningful abstracts (context) in the flow of human speech
• Analysis of the speaker’s emotional state
• Making summary of the human speech flow abstract to determine its subject or topic

Our human speech recognition technologies allow you to use them in automation projects dedicated to the agencies and firms from the governmental and commercial sectors of the economy.

We have developed a few software products to automate business communication with different representatives of business society: banking and finance, medicine, tourism and traveling, retail, and civil passenger transportation.

IVR using voice control

This is a software product dedicated to forward incoming telephone calls received by an emergency center or office IP-telephony systems utilizing recognition technology of subscriber’s oral questions and answers. This technology is primarily based on techniques dedicated to human speech recognition.

The software platform can be used to fulfill both simple and sophisticated scenarios, either intended to determine a subject of the request and to forward the subscriber to a person in charge, or intended to guide the subscriber to a relevant piece of information:

• objects GPS search and location
• automatic ticketing based on information provided by the speaker
• providing information services

MIR — human speech recognition platform

MIR IT platform designed for human speech recognition can be used as:

• a cloud-based IT-service and payment per requested information.
• recognition accuracy of the general model comprises 84%.
• on-premises IT solution to be installed and used on the client’s servers.

You can purchase Professional Service to build IT solutions specifically dedicated to meet the client’s needs.

For the client’s convenience and to ease the commissioning, we have authored a detailed manual. It helps our clients to easily integrate our IT solution in their software products.

MIR IT platform is provided with SDK to be ready for use in iOS and Android. It facilitates the use of our IT solution as an integral part of different mobile apps.

MIR IT platform can support industry-specific standards. It is compatible with major IP-telephony platforms since it utilizes MRCP protocol.

MIR analytics

This is a software product dedicated to perform analysis of human speech flow to determine meaningful abstracts (subjects, facts and their logical connections).

MIR Analysis can perform the following functions:

• calculation of verbal activities parameters
• semantic parsing and interpretation of live human speech
• classification of the requests’ subjects to be able to prepare an automatic reply

All the requests shall be classified automatically. The system provides a reporting mechanism based on the request classification.

The user can do the following operations:

• Leverage statistical data concerning the requests received by an emergency contact center
• Become aware of the problematic areas and trends
• Ability to check and control quality of operators’ work.