Speech Translation Non-Verbal Information Preservation

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Speech translation technologies often produce translation results that differ from the speaker's intention due to the loss of non-verbal information during the conversion process, leading to inaccuracies in conveying the intended meaning.

Innovation Solution

The method involves receiving a first language-based speech signal, converting it into text including non-verbal information through voice recognition, and translating this text into a second language, considering probability information and non-verbal cues such as pauses and hesitation words to improve the accuracy of the translation.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Loss of information

If speech signal undergoes several conversion steps (voice recognition to translation), then translation is achieved, but translation accuracy deteriorates due to loss of non-verbal information

Engineering Contradiction:

Improvenon-verbal informationVSAvoidtranslation accuracy

Core Design Contradiction:

Loss of informationVSMeasurement precision

Solution Approach 1:

The patent extracts and preserves non-verbal information (pauses, hesitation words, stress patterns) from the speech signal before the translation process begins. This preliminary extraction ensures that crucial contextual cues are captured and maintained throughout subsequent conversion steps, preventing information loss that would otherwise occur during voice recognition and translation processing

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The patent introduces an intermediary processing layer that handles non-verbal information separately from the main translation pipeline. This intermediary component analyzes speech characteristics like pauses and hesitation words, then integrates this information back into the translation process, acting as a mediator that preserves contextual meaning through multiple conversion steps

Inventive Principle:

Principle #24Intermediary (Mediator)

2Productivity

If traditional voice recognition and translation steps are used, then speech translation is achieved, but speaker's intention is lost

Engineering Contradiction:

Improvetranslation speedVSAvoidfaithfulness to speaker's intention

Core Design Contradiction:

ProductivityVSReliability

Solution Approach 1:

The patent segments the speech signal processing into distinct components: verbal content extraction, non-verbal information extraction (pauses, hesitations, stress), and integrated translation. This segmentation allows each component to be processed independently with appropriate methods, preserving the relationship between verbal and non-verbal cues while maintaining translation efficiency

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent implements feedback mechanisms where extracted non-verbal information is continuously fed back into the translation process to adjust and refine the translation output. This feedback loop ensures that the translation remains faithful to the speaker's intended meaning by incorporating contextual cues from speech patterns, pauses, and hesitation words throughout the translation process

Inventive Principle:

Principle #23Feedback

Data Source

PatentUS10614170B2Method of translating speech signal and electronic device employing the same

Publication Date: 2020.04.07 SAMSUNG ELECTRONICS CO LTD

AI summary

A method of translating a first language-based speech signal into a second language is provided. The method includes receiving the first language-based speech signal, converting the first language-based speech signal into a first language-based text including non-verbal information, by performing voice recognition on the first language-based speech signal, and translating the first language-based text into the second language, based on the non-verbal information.