Speech Recognition Phonetic Alignment and Path Weighting
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
General-purpose speech recognition systems fail to accurately recognize specialized vocabulary and sentences in specific application scenarios, such as medical terms or literature, resulting in inaccurate and meaningless output without providing sufficient options for developers or users to improve the recognition results.
Innovation Solution
A speech recognition system that connects to an external general-purpose system, comprising a specific application speech recognition module to convert speech signals into phonetic text, a comparison module to align phonetic texts based on pronunciation similarity, and an enhancement module to weight paths and constitute recognized text, improving accuracy by integrating phonetic text alignment and path weighting.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If a general-purpose speech recognition system is used, then the system can provide basic speech recognition functionality, but the recognition accuracy deteriorates in specific application scenarios such as medical terms or literature works
Solution Approach 1:
The system segments the speech recognition process into multiple independent modules: a general-purpose speech recognition module for basic recognition, a specific application speech recognition module for domain-specific terminology, a comparison module for aligning results, and an enhancement module for integrating outcomes. This segmentation allows each module to specialize in its function, improving overall accuracy for specific applications while maintaining general functionality.
2Loss of information
If a general-purpose speech recognition system is used, then the system structure remains simple, but the system cannot provide detailed information such as segmentation and word confidence
Solution Approach 1:
The system implements a nested structure where the specific application speech recognition module and comparison module are integrated within the broader speech recognition system. The enhancement module receives input from multiple sources and integrates them hierarchically, allowing detailed information to be generated and processed within the system while maintaining a unified interface for users.
3Measurement precision
If phonetic text alignment and path weighting are integrated, then speech recognition accuracy improves in specific scenarios, but the system complexity increases
Solution Approach 1:
The comparison module performs phonetic text alignment in advance by comparing phonetic transcripts from different recognition modules before final result generation. The enhancement module pre-calculates path weights based on alignment quality and confidence scores. These preliminary actions prepare the data for more accurate final recognition without adding complexity to the user interface or real-time processing bottlenecks.
Data Source
AI summary
A speech recognition system and method thereof are provided. The speech recognition system connects to an external general-purpose speech recognition system, and including a storage unit and a processing unit. The storage unit stores a specific application speech recognition module, a comparison module and an enhancement module. The specific application speech recognition module converts a speech signal into a first phonetic text. The general-purpose speech recognition system converts the speech signal into a written text. The comparison module receives the first phonetic text and the written text, converts the written text into a second phonetic text, and aligns the second phonetic text with the first phonetic text according to similarity of pronunciation to output a phonetic text alignment result. The enhancement module receives the phonetic text alignment result, and constructs with the written text and the first phonetic text after path weighting to form an outputting recognized text.