Voice Recognition Model Selection by Utterer Characteristics

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

General-purpose voice recognition models fail to accurately reflect the utterance intentions of users due to not accounting for unique characteristics, leading to incorrect recognition results and unintended operations in electronic devices.

Innovation Solution

An electronic apparatus and method that identifies user characteristics, such as gender, age, and pronunciation patterns, to select the most suitable voice recognition model from a set of models tailored to different user groups, adjusting recognition accuracy based on success or failure to improve matching between utterance intentions and recognition results.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If a general-purpose voice recognition model is used, then the system can handle unspecified utterers, but the recognition accuracy does not reflect the utterer's unique characteristics leading to incorrect recognition results

Engineering Contradiction:

Improveapplicability to unspecified utterersVSAvoidrecognition accuracy

Core Design Contradiction:

Adaptability or versatilityVSMeasurement precision

Solution Approach 1:

The patent segments the general voice recognition model into multiple specialized models, each trained on voice data from specific utterer groups (e.g., males, females, children, seniors). The system identifies the utterer's characteristics and selects the appropriate specialized model, thereby resolving the contradiction between handling unspecified utterers and achieving accurate recognition by dividing the general model into targeted segments.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent applies local quality by creating voice recognition models with specialized characteristics tailored to specific utterer groups. Each model has quality optimized for its target group (e.g., pronunciation patterns, accent characteristics), allowing the system to maintain high recognition accuracy for each segment while preserving overall versatility through model selection based on utterer identification.

Inventive Principle:

Principle #3Local quality

2Measurement precision

If multiple voice recognition models are provided for different utterer groups, then recognition accuracy for specific groups improves, but the complexity of the system increases

Engineering Contradiction:

Improverecognition accuracyVSAvoidsystem complexity

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The patent applies preliminary action by pre-identifying the utterer's characteristics (gender, age group) before voice recognition occurs. The system maintains a database of utterer profiles and pre-selects the appropriate voice recognition model based on these characteristics, thereby reducing real-time processing complexity while maintaining high recognition accuracy through proactive model selection.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The patent introduces an intermediary component - the utterer characteristics identification module - that acts as a mediator between the user's voice input and the multiple voice recognition models. This intermediary analyzes voice characteristics, determines the appropriate utterer group, and selects the corresponding model, thereby managing system complexity through a centralized selection mechanism rather than requiring complex interactions between multiple models simultaneously.

Inventive Principle:

Principle #24Intermediary (Mediator)

3Reliability

If the voice recognition model is adjusted based on recognition success or failure, then the suitability between utterance intention and recognition result improves, but the loss of information about original recognition accuracy occurs

Engineering Contradiction:

Improvesuitability between utterance intention and recognition resultVSAvoidoriginal recognition accuracy information

Core Design Contradiction:

ReliabilityVSLoss of information

Solution Approach 1:

The patent implements feedback by using recognition results to update and refine the voice recognition models. When recognition succeeds or fails, the system feeds this information back into the model training process, adjusting the models to improve future recognition accuracy. This feedback mechanism continuously improves reliability while preserving original accuracy information through structured data collection and analysis of recognition outcomes.

Inventive Principle:

Principle #23Feedback

Data Source

PatentUS11488607B2Electronic apparatus and control method thereof for adjusting voice recognition recognition accuracy

Publication Date: 2022.11.01 SAMSUNG ELECTRONICS CO LTD

AI summary

Disclosed is an electronic apparatus which identifies utterer characteristics of an uttered voice input received; identifies one utterer group among a plurality of utterer groups based on the identified utterer characteristics; outputs a recognition result among a plurality of recognition results of the uttered voice input based on a voice recognition model corresponding to the identified utterer group among a plurality of voice recognition models provided corresponding to the plurality of utterer groups, the plurality of recognition results being different in recognition accuracy from one another; identifies recognition success or failure in the uttered voice input with respect to the output recognition result; and changes a recognition accuracy of the output recognition result in the voice recognition model corresponding to the recognition success, based on the identified recognition success in the uttered voice input.