Document Image Binarization for Colored Table Cell Text Extraction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods struggle to accurately and efficiently generate binary images from document images with multi-colored text and varying background colors, leading to errors in text recognition and extraction due to similar intensity of foreground and background.
Innovation Solution
A method and system that utilize a processor to determine negative, inverse negative, HSV, and grayscale images, compute foreground and background mean values, categorize cells as dark or light based on these values, and apply thresholding techniques to generate a binary image.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If deep learning models are used to determine cells in table structure, then text extraction capability is improved, but computational complexity increases
Solution Approach 1:
The patent segments the image processing task into distinct stages: negative map generation, line detection for cell boundaries, foreground-background separation, and binary image generation. This segmentation replaces the monolithic deep learning approach with multiple simpler, specialized processing steps that collectively achieve the same goal with reduced computational complexity.
Solution Approach 2:
The patent introduces intermediate representations (negative map image, line-detected cell structures, foreground mask) as mediators between the input document image and the final extracted text. These intermediates break down the complex recognition task into manageable stages, each handling a specific aspect of the problem with simpler algorithms.
2Productivity
If traditional text recognition is applied to images with similar foreground and background intensity, then processing speed is maintained, but recognition accuracy deteriorates
Solution Approach 1:
The patent transforms the image from color space to negative map space, where the intensity relationships between foreground and background are inverted and enhanced. This color/intensity transformation makes previously similar regions distinctly different, enabling accurate segmentation and recognition while maintaining processing efficiency through direct pixel-level operations.
Solution Approach 2:
The patent changes the parameter space by computing negative maps (inverting intensity relationships) and detecting line structures to identify cell boundaries. These parameter transformations convert the difficult problem of distinguishing similar intensities into the easier problem of detecting structural boundaries and contrast differences.
3Device complexity
If binary thresholding is applied without considering cell categorization, then processing steps are reduced, but binary image quality deteriorates
Solution Approach 1:
The patent applies different processing strategies to different types of cells based on their local characteristics. By categorizing cells as dark or light based on foreground-background intensity relationships, the system selects appropriate thresholding approaches for each cell type, optimizing binary image quality for heterogeneous document contents rather than using a uniform approach.
Data Source
AI summary
A method and system for generating binary image of a document image is disclosed. The method includes determining a negative map image, an inverse negative map image, an HSV image and a grayscale image of the document image. One or more cells corresponding to at least one table are detected based on detection of lines in the negative map image. For each of the cells, a foreground mean value, a background mean value, and a background mean HSV value is determined. Each of the cells are categorized as a dark cell or a light cell based on the foreground mean value and the background mean value. Contrast value of each cell is determined based on the foreground mean value and the background mean value. The binary image is determined based on the contrast value, a pre-defined threshold value, the background mean HSV value and the categorization of the corresponding cell.