Voice Recognition Correction System Using Association Degree
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional voice recognition systems face difficulties in real-time correction of voice recognition results, leading to discrepancies and reduced legibility due to semantic and time-series factors, requiring manual intervention by editors.
Innovation Solution
A correction system comprising an interface, calculator, and display controller that estimates parts of the voice recognition result to be corrected, calculates the degree of association between correction information and the result, and generates corrected display information using appropriate formats, allowing for real-time correction and improved legibility.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual correction is performed by editors following conventional steps, then correction accuracy can be improved, but correction time and operational complexity increase significantly
Solution Approach 1:
The system enables self-service correction by automatically analyzing voice recognition results, estimating correction needs, and generating corrected text without requiring manual editor intervention for each correction step
Solution Approach 2:
The system performs preliminary analysis of the voice recognition result to estimate parts that need correction before the actual correction process, preparing correction information in advance to streamline the subsequent correction operations
2Measurement precision
If multiple correction steps are required for real-time caption correction, then correction precision can be maintained, but device complexity and ease of operation deteriorate
Solution Approach 1:
The system merges multiple correction steps into a single integrated process by combining result analysis, correction estimation, and correction information generation into one automated workflow, reducing the number of separate operations required
Solution Approach 2:
The system introduces an intermediary analysis component that automatically processes the voice recognition result and generates correction information, acting as a mediator between the raw recognition output and the final corrected text
3Reliability
If conventional correction methods are used, then correction completeness can be achieved, but legibility and viewer understanding deteriorate due to semantic discrepancies
Solution Approach 1:
The system incorporates feedback mechanisms that analyze the semantic relationship between the original voice recognition result and the corrected text, ensuring that corrections maintain or improve viewer understanding rather than creating new discrepancies
Data Source
AI summary
A correction system of the embodiment includes an interface system, a calculator, a generator, and a display controller. The interface system receives correction information for correcting a voice recognition result. The calculator estimates a part of the voice recognition result to be corrected and calculates a degree of association between the part to be corrected and the correction information. The generator generates corrected display information comprising at least one of the correction information and the part to be corrected using a display format corresponding to the degree of association. The display controller outputs the corrected display information on a display.