Image Display Device Speech Query Handling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional speech recognition methods are complex and limited by hardware, leading to difficulties in understanding natural language queries and experiencing network interruptions, which hinder the performance of speech recognition services, especially on mobile devices with resource constraints.
Innovation Solution
An image display device that acquires speech queries, generates a list of candidate queries with similar semantics, and performs operations based on user selections, using both internal and external engines to ensure continuous service even during network interruptions, and adapts Q/A sets based on user context and usage patterns.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional speech recognition methods are used, then speech recognition functionality is achieved, but the system complexity and hardware requirements increase
Solution Approach 1:
The patent divides the speech recognition system into multiple independent modules: speech acquisition module, speech recognition module, query list generation module, and operation execution module. Each module performs a specific function, allowing the system to maintain complexity while improving reliability through modular architecture that can operate independently or in combination.
Solution Approach 2:
The system pre-generates query lists with candidate queries and their corresponding operations before user input is processed. This preliminary preparation of possible queries and operations reduces the computational burden during real-time speech recognition and improves service availability by having ready-made response options.
2Adaptability or versatility
If a complete Q/A service is implemented, then comprehensive query understanding is achieved, but the response time increases due to network interruptions
Solution Approach 1:
Instead of implementing a complete Q/A service that requires full network communication, the system performs partial action by maintaining a local cache of candidate queries and operations. This allows the device to respond to speech queries using locally stored information, reducing response time and eliminating network interruption delays while still providing versatile query understanding.
Solution Approach 2:
The patent introduces a local query cache as an intermediary between the speech recognition module and the network-based Q/A service. This intermediary stores pre-fetched candidate queries and operations, allowing the system to provide comprehensive query understanding capabilities while bypassing network delays through local lookup and selection.
3Adaptability or versatility
If cloud computing techniques are used for Q/A services, then comprehensive query processing is achieved, but network interruptions cause service failures
Solution Approach 1:
The system implements local quality by maintaining a distributed cache of candidate queries and operations on each device rather than relying solely on centralized cloud processing. This local storage capability ensures that query processing can continue offline or during network interruptions, improving service continuity while maintaining comprehensive processing capability when available.
Solution Approach 2:
The patent applies beforehand cushioning by pre-loading and caching candidate queries and operations locally on devices before network interruptions occur. This preparatory action creates a buffer that allows the system to continue providing query processing services during network outages, ensuring service continuity without sacrificing comprehensive processing capability.
4Measurement precision
If speech recognition processes all linguistic analysis steps, then accurate semantic extraction is achieved, but the processing complexity increases
Solution Approach 1:
The system performs preliminary action by pre-processing and categorizing speech queries into candidate query templates before full semantic analysis. This preliminary classification reduces the complexity of subsequent linguistic analysis by narrowing down the scope of processing to predefined query patterns, while still achieving accurate semantic extraction for the selected candidate.
Solution Approach 2:
Instead of applying all linguistic analysis steps to every speech input, the system uses partial action by selectively applying full semantic analysis only to the selected candidate query from the generated list. This reduces overall processing complexity while maintaining high semantic extraction accuracy for the final chosen query through focused analysis.
Data Source
AI summary
An image display device, a method for driving the same, and a computer readable recording medium are provided. The image display device includes a speech acquirer configured to acquire a speech query associated with a query created by a user, a display configured to display a query list composed of candidate queries having the same as or similar semantic as the acquired speech query, and an operation performer configured to perform an operation related to the query selected from the displayed query list.