Offline Speech Recognition via Local Voiceprint Matching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current speech recognition systems require a network connection to function, which hinders their operation when internet connectivity is poor or unavailable, as they rely solely on remote servers for data processing.
Innovation Solution
Implementing a local device with both a voice database and a result database that allows for offline analysis by comparing received voice data against stored history data, enabling speech recognition even without internet access by executing an off-line analysis when internet connectivity is lost.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If speech recognition is executed via remote server through network connection, then recognition accuracy can be improved, but the system cannot operate when network connection is bad or unavailable
Solution Approach 1:
The speech recognition system is segmented into two independent parts: online recognition (remote server) and offline recognition (local device). The local device maintains voiceprint templates and can perform independent recognition without network dependency, while the remote server provides enhanced recognition when available. This segmentation resolves the contradiction by making the system partially independent of network connections.
Solution Approach 2:
The local device performs preliminary action by pre-storing voiceprint templates and recognition data before network disconnection occurs. This allows the offline recognition function to operate immediately when the network is unavailable, ensuring continuous system availability without requiring real-time network connection.
2Measurement precision
If only online speech recognition is implemented, then recognition precision can be maintained, but the device complexity increases due to network requirements
Solution Approach 1:
The complex remote server recognition system is extracted from the local device, with only the essential voiceprint templates and offline recognition capability remaining locally. This extraction reduces local device complexity while maintaining recognition accuracy through the simplified offline model, and allows online enhancement when needed.
3Reliability
If offline voiceprint comparison is performed locally, then network connection requirement is eliminated, but recognition speed may be reduced compared to server-based processing
Solution Approach 1:
The recognition system uses parameter changes by maintaining pre-computed voiceprint templates with optimized feature representations. These templates are designed for fast local comparison, balancing the trade-off between offline capability and processing speed. The templates contain essential acoustic features that enable rapid matching while maintaining acceptable recognition accuracy.
Data Source
AI summary
A speech recognition method, an electronic device and a speech recognition system are provided. When a local device is not connected to Internet is determined, a voiceprint comparison between the received voice data and the history voice data stored in the voice database is executed to obtain the corresponding history voice data, and an associated history text data is found from a result database of the local device according to the obtained history voice data.