Speech sound identification method

Abstract

It is one of the objectives of the inventive method to reduce the burden of search within the set of possible candidates for a speech phrase to be recognized. The objective is achieved by a method in which regions of high complexity of a recognition grammar are paired with regions of lower perplexity by using additional constraints. The search proceeds then by evaluating the low-perplexity region of the grammar first finding word candidates which are used to limit the search effort that needs to be expended when recognizing the parts of speech corresponding to the higher perplexity part of the recognition grammar.

Description

technical field [0001] The invention relates to a method of recognizing speech which reduces the burden of searching among a set of possible recognition objects. Background technique [0002] Recently, devices and methods for automatically recognizing continuously narrated speech have become more and more important. Indeed, there are a wide range of fields, such as information services, customer support, etc., where a considerable amount of people-related costs can be avoided through the use of equipment that automatically responds to customer inquiries. [0003] The most important condition that devices and methods for automatic speech recognition must meet is that they must reliably recognize and understand the customer's speech input independently of specific narration conditions, such as speech speed, pitch, intelligibility, background noise, etc. [0004] There are many devices, such as automated telephone services, timetable information services, etc., which only work...

AI Technical Summary

Problems solved by technology

[0008] A major disadvantage of conventional methods and devices for large vocabulary speech recognition is the great complexity and large number of possible candidates for speech segments or elements to be retrieved and tested

Images