Voice Search Filtering via Repeated Input Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing voice recognition systems in smartphones inadvertently search and provide information for unintentional user voice inputs, leading to redundant and unwanted search results.
Innovation Solution
A method and apparatus that compares input voice data with a user-defined database, only initiating a search when the voice data is intentionally repeated, using a voice comparison unit and controller to determine if the input matches stored data and if a repeated input setting is enabled, thereby preventing unnecessary searches.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If voice recognition is continuously activated to search for information, then information corresponding to user voice is provided immediately, but unintentional voice inputs are also recognized and redundant search results are generated
Solution Approach 1:
The system performs preliminary actions by storing user voice patterns and search preferences in advance. When voice input is detected, the system compares it against stored patterns to determine if it matches a searchable query, filtering out unintentional inputs before initiating search. This preliminary filtering resolves the contradiction by preparing the system ahead of time to quickly recognize valid queries while ignoring noise.
Solution Approach 2:
The system implements feedback mechanisms where search history and user responses are continuously monitored. When redundant searches are detected, the system adjusts its voice recognition sensitivity and filtering criteria based on this feedback, learning from past false positives to improve future discrimination between intentional and unintentional voice inputs.
2Ease of operation
If voice recognition is continuously activated to search for information, then information corresponding to user voice is provided immediately, but information search is redundantly performed creating noise for users
Solution Approach 1:
The system extracts and isolates the essential characteristic of intentional voice input by comparing against stored user voice patterns. By separating valid search queries from unintentional speech through pattern matching, the system maintains ease of operation for legitimate searches while removing information pollution caused by redundant results from accidental inputs.
Solution Approach 2:
The system dynamically changes recognition parameters such as confidence thresholds and matching criteria based on the context and frequency of voice inputs. When redundant searches are detected, the system adjusts these parameters to become more selective, thereby reducing information pollution while preserving ease of access to legitimate search functionality.
3Productivity
If voice data is searched upon single input, then quick response is achieved, but unintentional voice inputs are processed leading to unwanted search results
Solution Approach 1:
The system performs preliminary pattern matching against stored voice templates before triggering a search. This preliminary verification step ensures that only voice inputs matching known user speech patterns initiate searches, maintaining high productivity for valid queries while improving reliability by filtering out accidental inputs that don't match stored patterns.
Solution Approach 2:
The system uses feedback from search outcomes and user responses to refine its voice recognition reliability. When unintentional searches are detected, the system adjusts its triggering criteria and pattern matching sensitivity, learning from feedback to improve the accuracy of search triggers while maintaining efficient response times for legitimate queries.
Data Source
AI summary
The present invention relates to a method and an apparatus for information search and, specifically, to a method and an apparatus for information search using voice recognition, which, when voice data inputted through a microphone is present in a user-defined database and is repeatedly inputted, searches for information corresponding to the voice data and provides the information to a user. To this end, the method for information search using voice recognition according to the present invention comprises the steps of: when voice data is inputted through a microphone, comparing the inputted voice data with voice data stored in a user-defined voice database; when coincident voice data is present as a result of the comparison, determining whether a repeated voice data input is set; when the repeated voice data input is set as a result of the determination, ascertaining whether voice data coincident with the inputted voice data is present in an internal memory; and when the coincident voice data is present as a result of the ascertainment, outputting an event and requesting information corresponding to the voice data from a search server.