Method of providing intelligent voice recognition model for voice recognition device
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Voice recognition devices face accuracy limitations due to external factors such as ambient noise and howling, which current technologies are unable to effectively mitigate.
Innovation Solution
A method is provided to improve voice recognition accuracy by obtaining space type information about the placement area of a voice recognition device, extracting relevant feature information, and generating a tailored voice recognition model using this data, which includes using sensing devices to gather physical, acoustic, or image information and combining it to create a customized voice recognition model.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If a general voice recognition model is used, then the device can operate in various environments, but the voice recognition accuracy decreases in specific noisy environments
Solution Approach 1:
The voice recognition model is dynamically adapted to different placement environments by collecting acoustic characteristics data from the specific environment and adjusting the model parameters accordingly. This allows the system to transition from a static general model to a dynamic environment-specific model, resolving the contradiction between versatility and accuracy.
Solution Approach 2:
The acoustic characteristics of the placement environment (such as noise profiles, reverberation characteristics, and frequency responses) are used to modify the parameters of the voice recognition model. By changing the model parameters based on environmental parameters, the system achieves both environmental adaptability and improved recognition accuracy in specific conditions.
2Measurement precision
If the voice recognition model is customized for specific environments, then the voice recognition accuracy improves, but the device complexity increases
Solution Approach 1:
The system performs preliminary acoustic environment characterization by collecting data about the placement environment before deploying the voice recognition model. This preliminary action includes measuring background noise, reverberation, and other acoustic properties, which are then used to pre-adjust the model parameters, simplifying the overall customization process.
Solution Approach 2:
The voice recognition device automatically performs environmental acoustic characterization and model adaptation without requiring manual configuration or complex user intervention. The system self-adjusts by collecting acoustic data from its placement environment and automatically modifying its recognition model, thereby improving accuracy while maintaining simplicity.
3Measurement precision
If sensing devices are added to collect environmental data, then the voice recognition accuracy improves, but the device complexity and cost increase
Solution Approach 1:
Existing sensing devices in the voice recognition system (such as microphones and processors) are utilized for dual purposes: both for voice recognition and for collecting environmental acoustic characteristics. This multi-functionality approach allows the system to gather environmental data without adding dedicated sensing hardware, thereby improving accuracy while avoiding increased hardware complexity and cost.
Solution Approach 2:
The existing microphone array and signal processing capabilities of the voice recognition device are used to simultaneously perform voice capture and environmental acoustic characterization. The system self-services by using its own hardware resources to collect the data needed for model adaptation, eliminating the need for additional specialized sensing devices.
Data Source
AI summary
A method of providing an intelligent voice recognition model includes obtaining space type information about a placement area of the voice recognition device, extracting space feature information from the space type information; and generating a predetermined voice recognition model matched to the extracted space feature information. At least one device implementing the method of providing the intelligent voice recognition model may be associated with an artificial intelligence module, a unmanned aerial vehicle (UAV), a robot, an augmented reality (AR) device, a virtual reality (VR) device, devices related to 5G services, and the like.