Voice Recognition Server Integration Device

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

General-purpose speech recognition servers often fail to provide optimal performance for specific user-defined words, leading to differences in recognition accuracy between general-purpose and specialized servers, making it impractical to selectively use them based on speech content.

Innovation Solution

A system that preliminarily evaluates and integrates recognition results from both general-purpose and specialized speech recognition servers using a user-specific word list, weighting results to prioritize accuracy and response speed, with an intermediate module learning and preserving parameters for optimal integration.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If a specialized speech recognition server is constructed for specific user words, then recognition accuracy for specific words is improved, but recognition performance for general words deteriorates and development cost increases

Engineering Contradiction:
Improverecognition accuracy for specific wordsVSAvoidrecognition performance for general words
Core Design Contradiction:
Measurement precisionVSAdaptability or versatility

Solution Approach 1:

The patent combines multiple speech recognition servers (specialized servers for specific words and general-purpose servers for other words) into an integrated system. The integration server coordinates these servers to work together, selecting appropriate servers based on the recognition task to achieve both high accuracy for specific words and maintained performance for general words.

Inventive Principle:
Principle #5Merging (Combining)

Solution Approach 2:

The patent applies different recognition approaches to different word types: specialized recognition models are used locally for specific user words where high accuracy is critical, while general-purpose recognition models are used for other words. This localized optimization resolves the contradiction by tailoring the recognition approach to the specific requirements of each word category.

Inventive Principle:
Principle #3Local quality

2Adaptability or versatility

If a general-purpose speech recognition server is used, then recognition performance for general words is maintained, but recognition accuracy for specific user words deteriorates

Engineering Contradiction:
Improverecognition performance for general wordsVSAvoidrecognition accuracy for specific words
Core Design Contradiction:
Adaptability or versatilityVSMeasurement precision

Solution Approach 1:

The patent introduces an integration server as an intermediary between the user and the speech recognition servers. This intermediary server receives recognition requests, determines whether they involve specific user words, and routes them to appropriate specialized or general-purpose servers. This mediator resolves the contradiction by enabling selective use of specialized models without compromising general recognition capabilities.

Inventive Principle:
Principle #24Intermediary (Mediator)

3Measurement precision

If multiple speech recognition servers are integrated, then overall recognition accuracy is improved, but system complexity and processing time increase

Engineering Contradiction:
Improveoverall recognition accuracyVSAvoidsystem complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent performs preliminary classification of recognition requests to determine whether they involve specific user words before invoking multiple servers. This advance determination allows the system to selectively engage specialized servers only when necessary, reducing unnecessary processing complexity while maintaining high accuracy for relevant cases.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent segments the speech recognition task into different parts handled by different servers: specialized servers handle specific user words while general-purpose servers handle other words. This segmentation allows each server to focus on its strength, improving overall accuracy while managing system complexity through clear division of labor.

Inventive Principle:
Principle #1Segmentation

4Measurement precision

If specialized speech recognition server is developed, then recognition accuracy for specific words is improved, but development cost and time increase

Engineering Contradiction:
Improverecognition accuracy for specific wordsVSAvoiddevelopment cost
Core Design Contradiction:
Measurement precisionVSEase of manufacture

Solution Approach 1:

The patent creates a multi-functional integration server that can coordinate both specialized and general-purpose recognition servers. This integration server provides universal functionality for managing diverse recognition tasks, allowing the system to achieve specialized recognition accuracy without requiring separate dedicated systems for each word type, thereby reducing overall development cost.

Inventive Principle:
Principle #6Universality (Multi-functionality)

Data Source

PatentEP2838085B1Voice recognition server integration device and voice recognition server integration method
Publication Date: 2019.09.04 FAURECIA CLARION ELECTRONICS CO LTD
  • EP2838085B1 patent drawingFigure 1
  • EP2838085B1 patent drawingFigure 2
  • EP2838085B1 patent drawingFigure 3

AI summary

The speech recognition result through the general-purpose server and that through the specialized speech recognition server are integrated in an optimum manner, thereby, a speech recognition function least in errors in the end being provided. The specialized speech recognition server 108 is constructed with the words contained in the user dictionary data in use as well as the performance of the general-purpose speech recognition server 106 is preliminarily evaluated with such user dictionary data. Based on such evaluation result, information related to which recognition results through the specialized and general-purpose speech recognition servers are adopted and to how the adopted recognition results are weighted to obtain an optimum recognition result is preliminarily retained in the form of a database. Upon executing recognition, an optimum recognition result is obtained by comparing the recognition results through the specialized and general-purpose servers with the parameter for recognition result integration 118.