Speech Correction for Typed Input Using Audio Confidence Scoring
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Users face challenges in correcting text errors on electronic devices, particularly when the device is in a mobile or non-stationary position, due to small keyboards or touchscreens, and difficulty in positioning a text cursor for corrections.
Innovation Solution
A method and system that receives typed words, generates candidate words, translates an audio stream into text using these candidates, assigns confidence scores, and replaces typed words with selected candidates when the confidence score exceeds a threshold, enabling hands-free correction of errors.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If users manually correct text errors on mobile devices, then input accuracy can be maintained, but operation complexity and time consumption increase due to small keyboards and difficulty in positioning text cursor
Solution Approach 1:
The patent replaces the mechanical interaction of manually typing corrections with acoustic field interaction. Users speak the correct words into the device, and the speech recognition system automatically substitutes the spoken words to correct typographical errors, eliminating the need for manual cursor positioning and keyboard input for corrections.
Solution Approach 2:
The system performs self-correction by automatically detecting and fixing typographical errors using speech recognition. The device monitors the audio stream, compares spoken words with typed words, and automatically replaces incorrect typed words with the correct spoken words without requiring user intervention for each correction.
2Measurement precision
If users manually correct text errors on mobile devices, then input accuracy can be maintained, but time consumption increases due to the correction process
Solution Approach 1:
The speech recognition system operates continuously in the background, monitoring the audio stream alongside the text input process. Corrections are made in real-time as the user speaks, eliminating the need to pause typing, position cursor, and manually edit, thus maintaining continuous productive action.
Solution Approach 2:
The system generates candidate words based on the audio stream before the user completes their text input. By preparing corrections in advance based on spoken words, the system can immediately replace incorrect typed words without requiring the user to pause and manually locate errors, reducing correction time.
3Productivity
If speech recognition is used to correct typed words, then correction speed and accuracy improve, but device complexity increases due to additional components and processing
Solution Approach 1:
The speech recognition system serves multiple functions: it corrects typographical errors, provides autocomplete suggestions, and can transcribe spoken text to text. This multi-functionality justifies the added complexity by delivering multiple benefits from a single integrated system rather than requiring separate correction tools.
Solution Approach 2:
The patent combines the speech recognition module with the existing text input and processing systems. The audio stream processing, candidate word generation, and text replacement functions are integrated into the existing device architecture, sharing computational resources and processing pipelines to minimize the increase in overall system complexity.
4Measurement precision
If confidence scoring is implemented for candidate words, then input accuracy improves, but processing time and computational resources increase
Solution Approach 1:
The system calculates confidence scores for multiple candidate words generated from the audio stream, then applies a threshold filter to select only the most likely correct words. This partial evaluation approach processes only the necessary candidates rather than exhaustively analyzing all possible words, maintaining high accuracy while controlling processing time.
Data Source
AI summary
A method, computer program product, and system are provided for correcting one or more typed words on an electronic device. The method can include receiving one or more typed words from a text input device and generating one or more candidate words for the one or more typed words. The method can also include receiving an audio stream at the electronic device that corresponds to the one or more typed words. The audio stream can then be translated into text using the one or more candidate words, where the translation includes assigning a confidence score to each of the one or more candidate words. Based on the confidence score associated with each of the one or more candidate words, a candidate word can be selected among the one or more candidate words to represent each portion of the text. A word from the one or more typed words can be replaced with the selected candidate word based on the value of the confidence score associated with the selected candidate word.