Document Region Segmentation for Mixed Table and Text OCR
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing systems face difficulties in efficiently converting paper documents with both tables and character strings into specific file formats like .txt and .csv, as they often result in cumbersome handling and incorrect formatting when both elements are present in a single document.
Innovation Solution
An image processing apparatus with a character recognition unit and file generation unit that allows users to select specific file formats, enabling the generation of electronic files for either character strings or tables in formats like .txt or .csv, by analyzing and separating these elements within the document.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If OCR processing is applied to the entire document, then all content is converted to character codes, but the output format cannot distinguish between tables and character strings
Solution Approach 1:
The patent divides the document into distinct regions: table regions and character string regions. By segmenting the processing target, the system can apply different output formats to different regions, allowing tables to be exported as CSV and character strings as TXT, thus resolving the contradiction between processing efficiency and handling convenience.
Solution Approach 2:
The patent applies different processing qualities to different parts of the document. Table regions receive CSV format processing while character string regions receive TXT format processing. This local differentiation allows each region to be handled in the most appropriate format, improving both productivity and ease of operation.
2Device complexity
If a single file format is designated for the entire document, then processing is simple, but the output cannot be optimized for different content types
Solution Approach 1:
The patent introduces dynamic format selection based on content type. Instead of a static single format for the entire document, the system dynamically determines the appropriate format (CSV or TXT) for each region based on its content characteristics, achieving both simplicity and adaptability.
Solution Approach 2:
The patent creates a multi-functional processing system that can output multiple file formats (CSV and TXT) from a single processing operation. This universal approach allows the system to adapt to different content types without requiring separate processing steps, maintaining simplicity while enhancing format versatility.
3Productivity
If the entire document is processed as one region, then processing steps are reduced, but the output file becomes cumbersome to handle
Solution Approach 1:
The patent segments the document processing into distinct table region processing and character string region processing. This segmentation enables the generation of appropriately formatted files for each content type, improving file handling ease while maintaining processing efficiency through automated region identification and separate format application.
Data Source
AI summary
An image processing apparatus includes an image reading unit, a character recognition unit, an operation unit, and a file generation unit. The image reading unit generates image information which includes at least one of a character string and a table. The character recognition unit recognizes characters and generates character information corresponding to each character included in one of the at least one of the character string and the table. The operation unit receives, from a user, a selected file format of an electronic file to be generated. The file generation unit generates an electronic file including the generated character information in accordance with the selected file format. The character recognition unit recognizes the characters and generates the character information in the generated image information in accordance with the one of the at least one of the character string and the table that corresponds to the selected file format.