Image Layout Analysis via Coordinate Extraction

Solution Overview

Problem

Existing layout analysis techniques are algorithmically complex and computationally heavy, relying on direct image processing and semantic analysis to divide characters in images into paragraphs, which is inefficient for machine reading applications like converting paper books to electronic books or converting image characters to sound signals.

Innovation Solution

A computer-implemented layout analysis method that obtains coordinate information of text character lines, generates a layout model with simpler data structures indicating character presence, and performs paragraph division based on relative positional relationships, reducing computational load and maintaining precision.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If direct image processing and semantic analysis algorithms are used for layout analysis, then layout analysis precision can be maintained, but algorithm complexity and computational load increase significantly

Engineering Contradiction:

Improvelayout analysis precisionVSAvoidalgorithm complexity

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The patent segments the layout analysis process into two distinct stages: first extracting coordinate information of text character lines using simple image processing to obtain bounding boxes, then performing paragraph division based on relative positional relationships of these coordinates. This segmentation avoids the need for complex semantic analysis while maintaining precision by focusing on spatial relationships that define paragraph structure.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent extracts only the essential coordinate information (bounding box positions) of text character lines from the image, discarding unnecessary pixel-level details and semantic content. This extraction approach reduces computational load by working with simplified geometric data rather than full image data, while still providing sufficient information for accurate paragraph division.

Inventive Principle:

Principle #2Taking out (Extraction)

2Measurement precision

If direct image processing and semantic analysis algorithms are used for layout analysis, then layout analysis precision can be maintained, but computational load increases significantly

Engineering Contradiction:

Improvelayout analysis precisionVSAvoidcomputational load

Core Design Contradiction:

Measurement precisionVSUse of energy by moving object

Solution Approach 1:

The patent segments the layout analysis process into two distinct stages: first extracting coordinate information of text character lines using simple image processing to obtain bounding boxes, then performing paragraph division based on relative positional relationships of these coordinates. This segmentation avoids the need for complex semantic analysis while maintaining precision by focusing on spatial relationships that define paragraph structure.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent extracts only the essential coordinate information (bounding box positions) of text character lines from the image, discarding unnecessary pixel-level details and semantic content. This extraction approach reduces computational load by working with simplified geometric data rather than full image data, while still providing sufficient information for accurate paragraph division.

Inventive Principle:

Principle #2Taking out (Extraction)

3Measurement precision

If complex image processing algorithms are used for paragraph division, then paragraph division accuracy can be improved, but processing speed decreases

Engineering Contradiction:

Improveparagraph division accuracyVSAvoidprocessing speed

Core Design Contradiction:

Measurement precisionVSSpeed

Solution Approach 1:

The patent creates a simplified geometric representation (copy) of the text layout by extracting bounding box coordinates of character lines. This copy contains only the essential spatial information needed for paragraph division, eliminating the need to process the original complex image data repeatedly. The bounding box coordinates serve as a lightweight surrogate that preserves paragraph structure information while enabling faster computation.

Inventive Principle:

Principle #26Copying

Solution Approach 2:

The patent segments the layout analysis process into two distinct stages: first extracting coordinate information of text character lines using simple image processing to obtain bounding boxes, then performing paragraph division based on relative positional relationships of these coordinates. This segmentation avoids the need for complex semantic analysis while maintaining precision by focusing on spatial relationships that define paragraph structure.

Inventive Principle:

Principle #1Segmentation

Data Source

PatentUS10467466B1Layout analysis on image

Publication Date: 2019.11.05 NEXTVPU (SHANGHAI) CO LTD

AI summary

The present application relates to layout analysis on an image. The layout analysis method comprises: obtaining coordinate information of one or more character lines in an image; generating a layout model corresponding to the image by setting character data in regions in a data structure that correspond to the coordinate information of the one or more character lines, the character data including data indicating the presence of character; scanning the generated layout model to read the character data from the layout model, and performing paragraph division on the layout model based on a relative positional relationship of the read character data in the layout model.