Speech Recognition Keyword Filtering for User-Specific Information Extraction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing technologies fail to effectively extract and display information specific to a user from speech data during voice calls, particularly due to limitations in display area and the inability to extract keywords not pre-input by the user.
Innovation Solution
An information processing device equipped with speech recognition, filtering, and output means that generates a character string from speech data and filters keywords based on pre-stored words relevant to the speaker, allowing for the extraction and display of user-specific information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If all extracted keywords are displayed in the limited display area, then complete speech information is provided, but irrelevant keywords waste display space and reduce information value
Solution Approach 1:
The patent extracts and displays only keywords that match the user's profile from the speech recognition results, removing irrelevant keywords. This selective extraction resolves the contradiction by displaying complete relevant information while minimizing waste of limited display space.
Solution Approach 2:
The patent applies different treatment to different keywords based on their relevance to the user's profile. Keywords matching the profile are displayed with higher priority or prominence, while non-matching keywords are excluded or given lower priority, creating localized quality differentiation in the display.
2Adaptability or versatility
If keywords are selected based on user profile matching, then user-specific valuable information is prioritized, but keywords not pre-input by the user cannot be extracted
Solution Approach 1:
The patent performs preliminary registration of user profile keywords in advance, enabling the system to identify and extract relevant keywords during speech processing. This preliminary preparation allows the system to adaptively extract user-specific information while maintaining the capability to recognize new keywords through profile expansion mechanisms.
3Ease of operation
If manual note-taking is required during voice calls, then complete control over note content is achieved, but user convenience is significantly reduced
Solution Approach 1:
The patent implements automatic keyword extraction and display functionality that operates without requiring manual user intervention. The system automatically processes speech data, extracts relevant keywords based on user profile, and displays them, allowing users to passively receive important information during voice calls without manual note-taking effort.
Data Source
AI summary
Provided are an information processing device, etc. that is capable of extracting information specific to a user from speech data. This information processing device is provided with: speech recognition means for generating a character string based on speech data; filtering means for filtering one or more keywords extracted from the character string generated by the speech recognition means, based on one or more words which are relevant to a speaker of the speech data and stored in advance; and output means for outputting a result of the filtering performed by the filtering means.