Document Processing With Layout Segmentation for Smaller Files
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional document processing systems generate large PDF files due to treating each whole page as an image, even when only a small portion of the page contains data, leading to file size issues during transmission.
Innovation Solution
A document processing system that segments the layout of a page, identifies block types, converts text blocks to text, and stores meta information, constructing a formatted page to determine if it has a smaller storage size than the preprocessed image.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If the whole page is scanned and treated as an image, then complete page content is captured, but the file size becomes large
Solution Approach 1:
The patent segments the page layout into multiple blocks (text blocks, image blocks, table blocks, etc.) and processes each block separately. This allows the system to convert text blocks to text format for efficient storage while keeping image blocks as compressed images, thereby reducing overall file size while preserving complete page content.
Solution Approach 2:
The patent applies different processing methods to different parts of the page based on their content type. Text blocks are converted to text format with metadata, image blocks are compressed using image compression algorithms, and table blocks are converted to structured data. This localized processing optimizes storage efficiency for each content type while maintaining overall content completeness.
2Quantity of substance
If image compression is applied to the whole page, then file size is reduced, but the compression efficiency remains insufficient
Solution Approach 1:
The patent divides the page into distinct blocks and applies specialized compression methods to each block type. Text blocks use text encoding with metadata storage, image blocks use image-specific compression, and table blocks use structured data formats. This segmentation enables much higher compression efficiency compared to applying generic image compression to the entire page.
Solution Approach 2:
The patent changes the representation parameters of different content types. Text is converted from image pixels to text characters with associated metadata (font, size, position). Images are compressed using optimized image formats. Tables are converted to structured data with defined schemas. These parameter changes enable more efficient storage while preserving content quality.
Data Source
AI summary
Apparatuses and methods relate generally to reducing data usage by a document processing system. In one such method, a segmented layout of a page from a preprocessed image thereof is detected. Components of the layout are separated into one or more blocks. Each type of the one or more blocks are identified. A block of the one or more blocks identified as a text block is converted to text. Meta information for the block is stored. A formatted page for the page including the text and the meta information is constructed. It is determined whether the formatted page has a smaller storage size than the preprocessed image associated with the formatted page.