Speech Recognition Utterance-End Detection With Adaptive Hangover Time
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing electronic devices struggle to accurately determine the end of a user's utterance in speech recognition tasks, particularly when dealing with incomplete or ambiguous sentences, leading to inefficiencies in natural language processing.
Innovation Solution
An electronic device that adaptively adjusts a 'hangover time' for determining the end of an utterance based on the identified utterance type, which can be complete, incomplete, or ambiguous, using a deep-learning technique to analyze partial texts and user characteristics.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a fixed hangover time is used to determine the end of an utterance, then the device complexity is reduced, but the measurement precision of utterance end detection deteriorates
Solution Approach 1:
The patent implements dynamic hangover time adjustment by classifying utterances into different types (complete sentence, incomplete sentence, ambiguous sentence) and applying different hangover time values accordingly. The processor dynamically selects the hangover time based on the classified utterance type, replacing the fixed hangover time approach with an adaptive mechanism that improves detection precision without requiring complex real-time analysis.
2Reliability
If the hangover time is increased to ensure complete sentence recognition, then the reliability of speech recognition is improved, but the loss of time in processing increases
Solution Approach 1:
The patent applies different hangover time values to different utterance types: a longer hangover time is applied to incomplete and ambiguous sentences to ensure reliable recognition, while a shorter hangover time is applied to complete sentences to reduce processing delay. This localized quality adjustment ensures high reliability where needed while minimizing time loss in routine cases.
3Adaptability or versatility
If speech recognition processes all utterances with the same hangover time, then the ease of operation is maintained, but the adaptability to different utterance types deteriorates
Solution Approach 1:
The patent segments utterances into distinct types (complete sentence, incomplete sentence, ambiguous sentence) based on classification criteria, and applies different hangover time values to each segment type. This segmentation approach enables the system to adapt to different utterance types effectively while maintaining manageable complexity through clear classification rules.
Data Source
AI summary
Various embodiments of the disclosure relate to an apparatus for supporting automatic speech recognition in an electronic device capable of performing natural language processing, and an operating method thereof. To this end, based on one or more partial texts recognized sequentially from an audio signal input from a microphone, the electronic device may identify an utterance type for the input audio signal. The electronic device may adaptively determine a hangover time, based on the utterance type, and may detect an end of an utterance of the audio signal by applying the determined hangover time.