Voice Template Distinctness in Speech Recognition Training
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Voice templates created in voice-directed workflow systems often lack distinctness, leading to similarity with other words in the application vocabulary, which erodes speech recognition performance, causing errors and reducing productivity.
Innovation Solution
A method for dynamic training analysis that compares voice templates to other templates during creation, prompting users to adjust spoken words to make them less similar, using instructions such as enunciation prompts or alternative words to ensure distinctness.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If custom voice templates are created for each word, then speech recognition accuracy should improve, but template similarity between words increases leading to recognition errors
Solution Approach 1:
The system performs preliminary comparison of voice templates against existing templates in the vocabulary before finalizing template creation. This pre-check prevents creation of similar templates that would cause recognition errors, ensuring distinctness is established beforehand rather than corrected afterward.
Solution Approach 2:
The system provides feedback to the user when template similarity is detected, prompting them to create a new template with different pronunciation characteristics. This iterative feedback loop ensures the final template is sufficiently distinct from other vocabulary words while maintaining accurate representation of the intended word.
2Measurement precision
If workers create voice templates for all words, then speech recognition performance improves, but training time and productivity decrease
Solution Approach 1:
The system automatically identifies and creates templates only for words that are likely to cause confusion (such as phonetically similar words like 'five' and 'nine'), rather than requiring manual template creation for all vocabulary words. This localized approach maintains high recognition performance for critical words while significantly reducing overall training time.
3Reliability
If voice templates are made more distinct through user adjustment, then template discrimination improves, but training complexity and user burden increase
Solution Approach 1:
The system automatically compares created templates against the vocabulary and provides guidance prompts to users when similarity is detected. This self-service mechanism handles the complexity of template comparison and distinction automatically, requiring users only to make minor adjustments rather than navigating complex training procedures.
Data Source
AI summary
Speech recognition systems that use voice templates may create (or update) voice templates for a particular user by training (or re-training). If a training results in a vocabulary with similar voice templates, then the speech recognition system's performance may suffer. The present invention provides embraces methods for training a speech recognition system to prevent voice template similarity. In these methods, a trained word's voice template may be evaluated for similarity to other vocabulary templates prior to enrolling the voice template into the vocabulary. If template similarity is found, then a user may be prompted to retrain the system using an alternate word. Alternatively, the user may be prompted to retrain the system with the word spoken more clearly. This dynamic enrollment training analysis insures that all templates in the vocabulary are distinct.