Context-Aware Speech Recognition Thresholds for Emotion and Noise

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing voice and speech recognition systems are challenged by ambient noise and user voice variations, leading to reduced accuracy in different environments and emotional states.

Innovation Solution

Implementing a voice or speech recognition threshold based on contextual information and emotional classification, using trained models tailored for specific environments and emotional states to enhance recognition accuracy.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Speed

If voice recognition is performed in real-time during a conversation, then the response speed is improved, but the recognition accuracy deteriorates due to incomplete contextual information

Engineering Contradiction:
Improveresponse speedVSAvoidrecognition accuracy
Core Design Contradiction:
SpeedVSMeasurement precision

Solution Approach 1:

The system performs preliminary voice recognition on the first speaker's utterance before the second speaker responds. This preliminary recognition creates a preliminary result that is later corrected with full contextual information from both speakers, enabling fast initial response while maintaining accuracy through subsequent refinement

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The voice recognition system dynamically adjusts its processing based on conversation timing. It performs preliminary real-time recognition when the first speaker is talking, then updates and corrects the recognition results after the second speaker responds, creating a dynamic two-stage recognition process that adapts to conversation flow

Inventive Principle:
Principle #15Dynamics

2Productivity

If the voice recognition system processes only the current speaker's input, then the processing speed is improved, but the recognition accuracy deteriorates due to lack of contextual information from other speakers

Engineering Contradiction:
Improveprocessing speedVSAvoidrecognition accuracy
Core Design Contradiction:
ProductivityVSMeasurement precision

Solution Approach 1:

The system performs preliminary voice recognition on the first speaker's utterance before the second speaker responds. This preliminary recognition creates a preliminary result that is later corrected with full contextual information from both speakers, enabling fast initial response while maintaining accuracy through subsequent refinement

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The system uses the second speaker's response as feedback to correct and improve the preliminary recognition result. The recognition result generation unit updates the preliminary result by incorporating contextual information from both speakers, creating a feedback loop that enhances accuracy without sacrificing initial processing speed

Inventive Principle:
Principle #23Feedback

3Measurement precision

If the system collects and processes contextual information from multiple speakers, then the recognition accuracy is improved, but the device complexity increases

Engineering Contradiction:
Improverecognition accuracyVSAvoidsystem complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The system segments the voice recognition process into distinct units: a first voice recognition unit for the first speaker, a second voice recognition unit for the second speaker, and a recognition result generation unit that integrates both results. This segmentation allows independent processing of each speaker's input while maintaining overall system manageability

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The system performs preliminary voice recognition on the first speaker's utterance before the second speaker responds. This preliminary recognition creates a preliminary result that is later corrected with full contextual information from both speakers, enabling fast initial response while maintaining accuracy through subsequent refinement

Inventive Principle:
Principle #10Preliminary action

4Measurement precision

If the system waits for the complete conversation context before generating recognition results, then the recognition accuracy is improved, but the response time increases

Engineering Contradiction:
Improverecognition accuracyVSAvoidresponse time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The system performs preliminary voice recognition on the first speaker's utterance before the second speaker responds. This preliminary recognition creates a preliminary result that is later corrected with full contextual information from both speakers, enabling fast initial response while maintaining accuracy through subsequent refinement

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The voice recognition system dynamically adjusts its processing based on conversation timing. It performs preliminary real-time recognition when the first speaker is talking, then updates and corrects the recognition results after the second speaker responds, creating a dynamic two-stage recognition process that adapts to conversation flow

Inventive Principle:
Principle #15Dynamics

Data Source

PatentEP4377954B1Voice or speech recognition using contextual information and user emotion
Publication Date: 2026.05.06 QUALCOMM INC
  • EP4377954B1 patent drawingFigure 1A
  • EP4377954B1 patent drawingFigure 1B
  • EP4377954B1 patent drawingFigure 2

AI summary

A method of voice or speech recognition in varied environments and/or user emotional states executed by a processor of a computing device is provided. The method comprises: determining a voice or speech recognition threshold for voice or speech recognition based on information obtained from contextual information detected in an environment from which a received audio input was captured by the computing device and an emotional classification of a user's voice in the received audio input (310); determining a confidence score for one or more key words identified in the received audio input (312); and outputting results of a voice or speech recognition analysis of the received audio input in response to the determined confidence score exceeding the determined voice or speech recognition threshold (314).