Multilingual Voice Recognition for Entity Name Command Parsing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing voice recognition systems struggle to accurately recognize user voices in multiple languages, leading to inconsistent recognition results and user inconvenience due to the need to manually change settings or re-utter in the correct language.
Innovation Solution
An electronic apparatus that divides user voices by languages and applies different voice recognition models to identify entity names and generate control information, improving recognition accuracy and reducing processing time by recognizing and processing only unrecognized segments of the user's utterances in multiple languages.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If a single voice recognition model is used for multiple languages, then the device complexity is reduced, but the voice recognition accuracy deteriorates when the user speaks in a language other than the default setting
Solution Approach 1:
The system makes a single voice recognition model capable of handling multiple languages by integrating language detection and dynamic adaptation functions. The model automatically identifies the input language and adjusts its recognition parameters accordingly, making it universally applicable across different languages without requiring separate models for each language.
Solution Approach 2:
The system changes the operational parameters of the voice recognition model based on the detected language. By dynamically adjusting recognition thresholds, language-specific phoneme patterns, and processing parameters according to the identified language, the system maintains high accuracy across multiple languages while using a single model framework.
2Measurement precision
If multiple voice recognition models for different languages are maintained, then the voice recognition accuracy is improved, but the device complexity and processing time increase
Solution Approach 1:
The system segments the voice recognition process into distinct stages: language detection, model selection, and recognition execution. By dividing the overall process and maintaining only the necessary model components in memory while others are stored externally, the system achieves multi-language support without proportionally increasing device complexity.
Solution Approach 2:
The system introduces an external server as an intermediary to manage multiple language models. The local device maintains a lightweight client that communicates with the server for model retrieval and processing, offloading the complexity of managing multiple full-scale voice recognition models from the device itself.
3Measurement precision
If the user manually changes voice recognition settings to match their spoken language, then the voice recognition accuracy is improved, but the ease of operation deteriorates due to the need to change settings or re-utter
Solution Approach 1:
The system performs language detection and model selection automatically without requiring user intervention. The voice recognition system serves itself by autonomously identifying the input language and configuring the appropriate recognition parameters, eliminating the need for users to manually change settings or re-utter their commands.
Solution Approach 2:
The system implements a feedback loop where the detected language information is fed back into the recognition process to automatically adjust settings. This closed-loop control ensures that the voice recognition system continuously adapts to the user's spoken language in real-time, maintaining high accuracy without requiring manual user input for configuration.
4Loss of time
If voice recognition is performed in a single default language, then the processing time is reduced, but the adaptability to multiple languages deteriorates
Solution Approach 1:
The system dynamically adapts its processing approach based on the detected language. Rather than using a static single-language model, the system flexibly adjusts its recognition parameters, model selection, and processing strategies in real-time according to the input language, maintaining efficient processing speeds across multiple languages.
Solution Approach 2:
The system applies language-specific optimization to different segments of the recognition process. By tailoring the recognition parameters and model characteristics to match the specific requirements of each detected language, the system achieves both fast processing and high adaptability, rather than using a one-size-fits-all approach.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Disclosed is an electronic apparatus. The electronic apparatus includes a memory configured to store first voice recognition information related to a first language and second voice recognition information related to a second language, and a processor to obtain a first text corresponding to a user voice that is received on the basis of first voice recognition information, based on an entity name being included in the user voice according to the obtained first text, identify a segment in the user voice in which the entity name is included, and obtain a second text corresponding to the identified segment of the user voice on the basis of the second voice recognition information, and obtain control information corresponding to the user voice on the basis of the first text and the second text.