Foreign Word Grammar for Speech Recognition Accuracy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing speech recognition systems face challenges in accurately recognizing spoken foreign words due to mispronunciation and the difficulty in recognizing spoken letters, especially in languages like English where letters with similar sounds exist, leading to low accuracy rates.
Innovation Solution
A system and method utilizing a foreign word grammar that includes rules for both spoken and spelling forms of foreign words, incorporating statistical weights and probabilities based on frequency of use, and utilizing a prefix tree with probability-incorporated transition arcs to improve recognition accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional speech recognition is used for foreign words, then the system can process any input, but the recognition accuracy is low due to mispronunciation and similar-sounding letters
Solution Approach 1:
The system changes the parameters of speech recognition by incorporating statistical weights and probabilities for different foreign words and their pronunciations. The prefix tree structure with probability annotations allows the system to dynamically adjust recognition thresholds and select the most likely word based on statistical data, thereby improving accuracy while maintaining the ability to handle various inputs.
Solution Approach 2:
The patent introduces an intermediary layer between speech input and word recognition: the prefix tree with statistical weights and probabilities. This intermediary structure acts as a mediator that filters and evaluates possible word matches, using pre-computed statistical data to guide the recognition process and resolve ambiguities caused by mispronunciation or similar-sounding letters.
2Measurement precision
If the system accepts any speech input for foreign words, then it can handle diverse pronunciations, but it cannot distinguish between legitimate words and random letter combinations
Solution Approach 1:
The system performs preliminary action by pre-computing and storing statistical weights and probabilities for legitimate foreign words in a prefix tree structure before actual speech recognition occurs. This advance preparation creates a framework of expected valid words and their likelihoods, enabling the system to quickly validate whether a recognized sequence is a legitimate word or random letters during the recognition process.
3Adaptability or versatility
If the system uses a large vocabulary of foreign words, then it can recognize more words, but the computational complexity and memory requirements increase
Solution Approach 1:
The patent segments the large vocabulary of foreign words into a prefix tree structure, where words are organized hierarchically by their prefixes. This segmentation allows the system to efficiently store and search through a large vocabulary by breaking it down into smaller, manageable prefix-based units, reducing both computational complexity and memory requirements compared to storing complete word lists.
Data Source
AI summary
Systems and methods are utilized for recognizing speech that is partially in a foreign language. The systems and methods receive speech input from a user and detect if a rule or sentence entry grammar structure utilizing a foreign word has been uttered. To recognize the foreign word, a foreign word grammar is utilized. The foreign word grammar includes rules for recognizing the uttered foreign word. Two rules may be included in the foreign word grammar for each legitimate or slang term included in the foreign word grammar. A first rule corresponds to the spoken form of the foreign word, and the second rule corresponds to the spelling form of the foreign word. The foreign word grammar may also utilize a prefix tree. Upon recognizing the foreign word, the recognized foreign word may be sent to an application to retrieve the pronunciation, translation, or definition of the foreign word.