Character Recognition Using Dual Magnification Image Acquisition
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing character recognition techniques face challenges in accurately recognizing characters from document images captured from web browsers due to variations in font size and display magnification, leading to incorrect newline position insertion and reduced accuracy.
Innovation Solution
The proposed solution involves acquiring first and second image data at different display magnifications, performing character recognition on the second image data, and adjusting the newline positions based on the correspondence relationship between the two images to improve accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If character recognition is performed on image data captured from web browsers at original display magnification, then the processing speed is maintained, but the recognition accuracy deteriorates due to small font sizes and low resolution
Solution Approach 1:
The patent applies preliminary action by acquiring reference image data at the original display magnification before performing character recognition. This reference data is used to determine newline positions and adjust the recognition results later, preparing the necessary information in advance to enable accurate recognition without requiring high-resolution images throughout the entire process
Solution Approach 2:
The patent changes the display magnification parameter by acquiring image data at both the original display magnification and an enlarged display magnification. This parameter change allows the system to capture characters at higher resolution for accurate recognition while using the original magnification data for structural reference, thereby improving recognition accuracy without permanently increasing processing time
2Measurement precision
If image data is acquired at enlarged display magnification for better character recognition, then the recognition accuracy improves, but the newline position insertion becomes incorrect due to layout changes
Solution Approach 1:
The patent merges the advantages of both original and enlarged display magnification images by acquiring reference image data at the original magnification and character recognition image data at the enlarged magnification. The system combines the high-resolution character details from the enlarged image with the accurate layout and newline position information from the original image, achieving both accurate character recognition and correct newline insertion
Solution Approach 2:
The patent uses the reference image data acquired at the original display magnification as an intermediary to transfer newline position information to the character recognition results obtained from the enlarged image data. This intermediary reference data acts as a bridge that preserves the original document structure while enabling high-accuracy character recognition
3Adaptability or versatility
If screen capture is performed from web browsers, then the document display can be verified, but the character recognition accuracy deteriorates due to browser-specific rendering variations and expansion functions
Solution Approach 1:
The patent changes the image acquisition parameter by obtaining image data through multiple methods: both through browser rendering (to verify display compatibility) and through direct drawing processes that output image data at specified magnifications. This parameter change allows the system to use browser-rendered images for compatibility verification while using directly drawn images for accurate character recognition, thereby overcoming browser-specific rendering variations
Data Source
AI summary
An apparatus for character recognition executes a first process for acquiring first image data which is an image in which string data containing one or more characters is drawn at a first magnification through a drawing process that outputs image data acquired by drawing characters at a display magnification, executes a second process for acquiring second image data which is an image in which the string data is drawn at a second magnification larger than the first magnification through the drawing process, executes a third process for acquiring a recognition result including a character code of each of the characters in the string data drawn in the second image data through a character recognition process, and executes a fourth process for adjusting a newline position of the recognition result acquired from the second image data by using a newline position of the string data drawn in the first image data.