Document Editing Apparatus for Mixed Text and Image Layout Preservation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing document editing technologies fail to accurately store and convey the meaning of documents with mixed text and image areas, leading to disrupted character sequences and misrecognition of picture areas, which decreases the usability of recognized text data.
Innovation Solution
A document editing apparatus and method that identifies and separates picture areas from document images, stores them as image files, performs character recognition on the remaining text areas, and converts position information to maintain the original document structure, allowing for the creation of web documents that preserve the layout and readability of the original document.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If area analysis is performed to separate text and picture areas in a document image, then character recognition accuracy is improved, but picture areas may be misrecognized and character sequences may be disrupted
Solution Approach 1:
The patent extracts picture areas from the document image before performing character recognition. The image processing unit identifies and extracts picture areas based on image characteristics, then the character recognition unit performs recognition only on the remaining text areas. This prevents picture areas from being misrecognized as text while maintaining accurate character recognition in text areas.
2Productivity
If character recognition is performed on the entire document image including picture areas, then recognition speed is improved, but misrecognition of picture areas occurs and character sequences are disrupted
Solution Approach 1:
The patent segments the document image into picture areas and text areas before character recognition. The image processing unit divides the document image by identifying picture areas and separating them from text areas. This segmentation allows character recognition to be performed efficiently on text areas without the computational overhead of analyzing picture areas, while maintaining accurate character sequences.
3Quantity of substance
If only text data is stored after recognizing text areas, then storage efficiency is improved, but the ability to store various media documents including pictures and graphs is decreased
Solution Approach 1:
The patent merges text data and picture area image data into a single web document structure. The web document creation unit combines the recognition results (text data with position information) and the extracted picture area images into an integrated HTML document. This allows the system to maintain storage efficiency while gaining the versatility to store and reproduce various media documents including text, pictures, graphs, and tables in their original layout.
Data Source
AI summary
A method for storing a document recognition result is proposed. The method includes selecting a picture area from a document image, storing an image of the selected picture area in an image file format, removing the selected picture area, filling the removed picture area with a surrounding background color, and performing character recognition of a text area.