Dynamic Text Weighting for Voice Recognition in Mobile Terminals
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current display devices lack effective voice recognition capabilities, leading to low recognition rates, especially in diverse and dynamic content environments like IPTV, where users need to interact with text-based content without manual PC operation.
Innovation Solution
A display device with a sound sensing unit, database, and controller that extracts and updates text weights based on user voice recognition, allowing for improved matching scores and enhanced interaction by storing and prioritizing text based on user input, and integrating network interfaces for search results and voice recognition servers.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If text is stored in a database for voice recognition, then voice recognition capability is enabled, but recognition accuracy remains low without dynamic weight updates
Solution Approach 1:
The patent applies dynamics by transitioning from a static database to a dynamic database that automatically updates text weights based on display frequency and user interaction. The controller continuously monitors displayed text and adjusts weights in real-time, making the voice recognition system adaptive to changing content and user preferences, thereby improving recognition accuracy and success rate simultaneously
Solution Approach 2:
The patent implements feedback by using the display unit as a feedback channel to the database. When text is displayed on the screen, the controller detects this and automatically updates the corresponding text weight in the database. This closed-loop feedback mechanism ensures that frequently displayed text (which is more likely to be spoken by users) receives higher weights, improving voice recognition accuracy without requiring manual intervention
2Adaptability or versatility
If database stores all displayed text, then voice recognition coverage is comprehensive, but database size and processing complexity increase
Solution Approach 1:
The patent applies parameter changes by introducing a weight parameter that quantifies the importance of each text entry. Instead of treating all text equally, the system assigns dynamic weights based on display frequency and timing. This parameter transformation allows the database to maintain comprehensive coverage while simplifying processing through weighted prioritization, reducing the effective search space during voice recognition
Solution Approach 2:
The patent implements preliminary action by pre-processing and weighting text data during display, so that when voice recognition is needed, the system already has prioritized candidates ready. The database is continuously updated with weighted text information in the background, preparing the recognition system in advance rather than processing everything in real-time when a voice command is given, thus reducing processing complexity
3Ease of operation
If voice recognition uses static text database, then system operation is simple, but recognition rate is low in dynamic content environments
Solution Approach 1:
The patent applies self-service by enabling the database to automatically update itself based on display content without external intervention. The controller monitors the display unit and autonomously adjusts text weights in the database based on what is currently shown and user interactions. This self-updating mechanism maintains high recognition accuracy in dynamic environments while keeping the user interface simple and requiring no manual database management
Solution Approach 2:
The system performs preliminary weighting and prioritization of text data during normal display operations, preparing the database in advance for accurate voice recognition. By continuously pre-processing text weights based on display frequency and timing, the system ensures high recognition rates are ready when needed, maintaining both simplicity and precision
Data Source
AI summary
A display device is disclosed. The display device comprises a display unit, a sound sensing unit receiving a user's voice, a database storing text displayed on the display unit for a predetermined time period, and a controller extracting from the database at least one text corresponding to a user's voice received within a predetermined time period.