Speech Recognition Using Partial Utterance Matching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional speech recognition systems fail to accurately recognize entire text from partial utterances, leading to hindered usability and incomplete function execution, as they rely solely on exact matches with candidate words or sentences.
Innovation Solution
An electronic device and speech recognition method that convert speech signals into text, determine the highest matching text sets based on word ratios and order, allowing for partial phrase recognition and execution of intended functions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If conventional speech recognition requires exact match with candidate words or sentences, then recognition accuracy is maintained, but usability is hindered when users utter only partial text
Solution Approach 1:
The patent applies partial action by allowing speech recognition to succeed with only a portion of the complete candidate text being uttered. The system calculates a coincidence ratio between the uttered speech and the candidate text, and determines that recognition is successful when this ratio exceeds a predetermined threshold, rather than requiring complete text matching.
Solution Approach 2:
The patent changes the recognition parameter from binary (exact match or no match) to a continuous coincidence ratio measurement. By introducing this ratio parameter and comparing it against a threshold, the system enables flexible recognition that adapts to partial utterances while maintaining controlled accuracy through the threshold mechanism.
2Productivity
If speech recognition waits for complete text utterance, then full context is captured, but recognition speed and user responsiveness are reduced
Solution Approach 1:
The patent applies preliminary action by performing recognition determination during the speech utterance process itself, rather than waiting for completion. The system continuously monitors the coincidence ratio as speech is being uttered and can determine recognition success mid-utterance, enabling faster response while still capturing sufficient contextual information through the ratio threshold mechanism.
3Reliability
If conventional systems only operate when final result coincides with candidate text, then false positives are minimized, but post-processing opportunities are lost
Solution Approach 1:
The patent enables post-processing by allowing recognition to proceed when the coincidence ratio exceeds the threshold, even if it's not a perfect match. This partial matching approach provides more opportunities for beneficial post-processing operations while the threshold mechanism continues to filter out clearly incorrect matches, maintaining reliability.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
An electronic device is disclosed. The electronic device comprises: a microphone for receiving voice; a memory for storing a plurality of text sets; and a processor for converting the voice, received via the microphone, into text, searching for words common to the converted text with respect to each of the plurality of text sets, and determining at least one text set of the plurality of text sets on the basis of the ratio of the searched common words.