Speech Summary Generation Through Contextual Word Extraction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing voice recognition systems for generating summaries rely heavily on dictionaries, leading to potential inaccuracies in reflecting the actual speech content.
Innovation Solution
A data generation device that includes an acquisition unit for voice data, a recognition unit for text generation, an extraction unit to select words based on predetermined conditions and ranges, and a generation unit to create summaries using these selected words.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If a dictionary-based approach is used for text selection in voice recognition summarization, then the summarization process becomes systematic and automated, but the accuracy of reflecting actual speech content deteriorates
Solution Approach 1:
The patent extracts key information directly from the voice data through voice recognition without relying on pre-defined dictionary constraints. The system identifies and extracts relevant text segments that accurately reflect the spoken content, removing the limiting factor of dictionary-based selection that previously caused accuracy deterioration.
Solution Approach 2:
The patent changes the selection criteria parameter from dictionary-matching to voice recognition-based text extraction. By altering the fundamental parameter of how text is selected (from static dictionary lookup to dynamic voice recognition output), the system maintains automation while significantly improving accuracy in reflecting actual speech content.
2Measurement precision
If voice recognition is performed without dictionary constraints, then speech content accuracy improves, but system complexity increases
Solution Approach 1:
The patent employs a voice recognition unit that serves multiple functions: it performs both the recognition of speech content and the selection of relevant text segments. This multi-functional approach eliminates the need for separate dictionary-matching modules, reducing overall system complexity while maintaining high accuracy in reflecting speech content.
Solution Approach 2:
The voice recognition system operates autonomously to generate summaries without requiring external dictionary resources. The system self-services by using its own recognition output as the basis for text selection and summary generation, thereby simplifying the system architecture while improving accuracy.
Data Source
AI summary
A data generation device includes an acquisition unit for acquiring voice data of speech, a recognition unit for generating text data by performing voice recognition on the voice data, an extraction unit for extracting, as an extracted word, a word satisfying a predetermined condition, from among a plurality of words included in the text data, and a generation unit for generating summary data indicating a summary of content of the voice data, by using the extracted word and a word that is within a predetermined range from the extracted word, from among the plurality of words included in the text data.