Proofreading System Using Character Image Similarity for Error Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current proofreading methods struggle to accurately detect writing errors in characters that are difficult for humans to distinguish at a glance, such as the alphabet 'T' and the Greek character 'Τ', leading to decreased search comprehensiveness.
Innovation Solution
A proofreading system that includes a dividing unit, an appearance frequency obtaining unit, an image generation unit, a similarity degree obtaining unit, and a presentation unit, which divides sentences into terms, images terms with low appearance frequencies, compares these images for similarity, and presents potential errors based on high similarity degrees, utilizing machine learning models for enhanced accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If words are searched using text-based search methods, then the search process is simple, but writing errors in similar characters (e.g., alphabet 'T' vs. Greek character 'Τ') cannot be detected
Solution Approach 1:
The patent converts character strings into visual images (character images) to enable visual comparison. By creating an image of the searched character and comparing it with images of candidate characters, the system can detect writing errors between visually similar but different characters, such as the alphabet 'T' and the Greek character 'Τ', which text-based search methods cannot distinguish.
Solution Approach 2:
The patent changes the representation parameter from text-based character codes to visual image parameters. By transforming characters into their visual representations and using image processing techniques, the system can detect errors based on visual similarity rather than just character code differences, thereby improving error detection accuracy for similar-looking characters.
2Measurement precision
If visual comparison of characters is performed, then similar but different characters can be detected, but the processing time and computational resources increase
Solution Approach 1:
The patent pre-processes and stores character images in a database before actual proofreading operations. By preparing character images in advance and organizing them in a database, the system can quickly retrieve and compare images during proofreading without performing time-consuming image generation and processing in real-time, thus reducing overall processing time while maintaining high distinction accuracy.
3Quantity of substance
If all characters are processed equally, then complete coverage is achieved, but processing efficiency decreases due to low-frequency characters
Solution Approach 1:
The patent applies different processing strategies to different characters based on their frequency and characteristics. By identifying characters with low appearance frequencies or high similarity to other characters, the system can prioritize these for detailed visual comparison while using faster text-based methods for common characters, thereby improving overall proofreading efficiency without sacrificing coverage of important characters.
Data Source
AI summary
A proofreading system that allows a user to easily judge whether or not there is an error in writing or the like. A proofreading method using a comparison image group obtained by dividing a sentence included in a comparison document group into a plurality of first terms and converting the first terms into images is provided. Specifically, first, a sentence included in a designated document is divided into a plurality of second terms, and the appearance frequency in the comparison document group of the plurality of second terms are obtained. Next, the second term with the appearance frequency lower than or equal to a threshold value of the plurality of second terms are imaged to obtain a verification image. After that, similarity degrees between the verification image and comparison images included in the comparison image group are obtained, and the first term represented by the comparison image with the highest similarity degree of the comparison images is presented. The presentation is performed by displaying that the second term represented by the verification image can be an error in writing of the first term represented by the comparison image having a high similarity degree with the verification image.