Handwritten Content Removal in Text Images Using Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods struggle to efficiently remove handwritten content from text images, leading to difficulties in maintaining confidentiality and issues with ink wastage and readability when printing text photos.
Innovation Solution
A method utilizing an image segmentation model to identify and blur handwritten pixels, creating a mask area, and then removing the handwritten content from the input image, optionally through pixel restoration or binarization processing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If handwritten content is removed from text images using conventional methods, then confidentiality is improved, but the removal efficiency is low and user convenience deteriorates
Solution Approach 1:
The patent replaces manual mechanical methods (manual erasing, cropping, or editing) with an automated image processing system that uses convolutional neural networks and pixel restoration algorithms to automatically detect and remove handwritten content, thereby improving user convenience while maintaining confidentiality
Solution Approach 2:
The system performs self-service by automatically detecting handwritten regions, generating masks, and restoring underlying printed text without requiring user intervention, making the confidentiality protection process autonomous and convenient
2Ease of manufacture
If text photos are printed directly without processing, then printing simplicity is improved, but ink wastage increases and readability deteriorates due to shadows
Solution Approach 1:
The patent applies preliminary shadow removal processing to text photos before printing, detecting and eliminating shadow regions in advance to prevent ink wastage and improve readability, while maintaining the simplicity of the printing process
3Measurement precision
If image segmentation model is used to identify handwritten pixels, then handwritten content detection precision is improved, but processing complexity increases
Solution Approach 1:
The patent segments the image processing task into distinct stages: handwritten region detection using CNN, mask generation through blurring operations, and pixel restoration using neighborhood algorithms. This segmentation allows each component to be optimized independently, managing complexity while maintaining high detection precision
Data Source
AI summary
Provided by the present disclosure are a method and device for removing handwritten content from a text image, and a storage medium. The method for removing handwritten content from a text image comprises: acquiring an input image of a text page to be processed, wherein the input image comprises a handwritten area, and the handwritten area comprises the handwritten content; using an image segmentation model to recognize the input image so as to obtain an initial handwritten pixel of the handwritten content; performing blurring processing on the initial handwritten pixel to obtain a handwritten pixel mask area; determining the handwritten content according to the handwritten pixel mask area; and removing the handwritten content from the input image to obtain an output image.