OCR Feedback System for Mobile Translation Accuracy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Optical character recognition (OCR) in mobile devices often produces errors due to inadequate image capture conditions, such as lighting, leading to incorrect or missing characters, which can result in low-quality machine language translations.
Innovation Solution
A computer-implemented method that receives an image, performs OCR, and if the translation score is below a threshold, provides user instructions to capture additional images with adjusted lighting or angles, combining images to improve OCR text accuracy, and selects alternative OCR texts to enhance translation quality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If OCR is performed on a single captured image, then the process is fast and simple, but the accuracy of OCR text and translation quality deteriorate due to inadequate image capture conditions
Solution Approach 1:
The system performs preliminary evaluation of the first image's quality and OCR/translation scores before committing to the final result. If scores are below thresholds, it proactively requests additional images rather than proceeding with potentially inaccurate results, thus ensuring accuracy while avoiding unnecessary complexity when the first image is sufficient
Solution Approach 2:
The system uses feedback from OCR scores and translation scores to determine whether additional images are needed. The quality assessment mechanism provides feedback about the sufficiency of the captured image, and this feedback drives the decision to either accept the first image or request additional captures, resolving the contradiction between speed and accuracy
2Measurement precision
If multiple images are captured to improve OCR accuracy, then translation quality improves, but the time required and user burden increase
Solution Approach 1:
The system evaluates the first image and determines in advance whether additional images are necessary based on predefined quality thresholds. This preliminary assessment prevents unnecessary additional captures, saving time when the first image is sufficient while ensuring accuracy is maintained when it is not
Solution Approach 2:
The system captures additional images only partially - specifically when and only when the quality metrics indicate insufficient accuracy. This avoids the excessive action of always capturing multiple images, thereby reducing time loss while still achieving the necessary translation accuracy when required
3Measurement precision
If the system requests additional images when translation score is low, then translation quality improves, but the number of processing steps increases
Solution Approach 1:
The system uses translation scores as feedback to automatically determine whether additional processing is needed. When scores exceed thresholds, the process terminates efficiently; when they fall below, the system triggers additional image capture and processing. This feedback-driven approach optimizes productivity by avoiding unnecessary steps while ensuring quality when needed
Solution Approach 2:
The quality assessment and threshold comparison are performed preliminarily on the first image before committing to extensive processing. This preliminary action filters out cases that don't need additional work, maintaining high productivity while ensuring that translation quality is improved only when genuinely necessary
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A computer-implemented technique includes techniques are presented for user image capture feedback for improved machine language translation. When machine language translation of OCR text obtained from an initial image has a low degree of likelihood of being an appropriate translation, these techniques provide for user image capture feedback to obtain additional images to obtain a modified OCR text, which can result in improved machine language translation results. Instead of user image capture feedback, the techniques may obtain the modified OCR text by selecting another possible OCR text from the initial OCR operation. In addition to additional image capturing, light source intensity and/or a quantity/number of light source flashes can be adjusted. After obtaining the modified OCR text, another machine language translation can be obtained and, if it has a high enough degree of likelihood, it can then be output to a user.