Automated Text Legibility Assessment in Digital Images

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Content providers face challenges in determining the legibility of text in digital images of pages, as low resolution images become difficult to read, and manual human review is time-consuming and costly for large repositories.

Innovation Solution

A computer-implemented method and system that performs text recognition to calculate text-to-page height ratios or word density, comparing these measures to thresholds to determine legibility, allowing for automated assessment and potential upgrading of image resolution for clear text display.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Quantity of substance

If image resolution is reduced to minimize storage costs, then storage requirements decrease, but text legibility deteriorates

Engineering Contradiction:
Improvestorage requirementsVSAvoidtext legibility
Core Design Contradiction:
Quantity of substanceVSMeasurement precision

Solution Approach 1:

The patent changes the parameter of image resolution to find an optimal balance point. By systematically varying resolution levels and evaluating text legibility at each level, the system identifies the minimum resolution required for acceptable legibility, thereby minimizing storage requirements while maintaining adequate text readability.

Inventive Principle:
Principle #35Parameter changes

Solution Approach 2:

The patent replaces manual human review of image legibility with an automated computer-based system. The automated system uses image processing algorithms to objectively measure text legibility metrics, eliminating the need for human readers to visually inspect each image and determine whether text is readable.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

2Measurement precision

If manual human review is used to determine text legibility, then legibility assessment accuracy improves, but processing time and cost increase

Engineering Contradiction:
Improvelegibility assessment accuracyVSAvoidprocessing time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The patent substitutes manual human review with an automated computer-based legibility assessment system. The automated system processes images using algorithms that measure text legibility objectively, achieving both high accuracy and rapid processing speeds, thereby eliminating the time loss associated with manual review while maintaining or improving assessment accuracy.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The patent enables the system to automatically assess its own output quality by implementing self-evaluation capabilities. The computer-based system independently measures text legibility in generated images without requiring external human verification, allowing the system to autonomously determine whether images meet legibility standards and make appropriate adjustments.

Inventive Principle:
Principle #25Self-service

Data Source

PatentUS7466875B1Method and system for determining the legibility of text in an image
Publication Date: 2008.12.16 NOKIA TECHNOLOGIES OY
  • US7466875B1 patent drawing
  • US7466875B1 patent drawing
  • US7466875B1 patent drawing

AI summary

Legibility of text in an image of a page is determined by comparing a measure of the text in the page image with a measure of the page image itself. In one aspect, a measure of the text in the page image may be the height of a line of text, while the measure derived from the page image may be the height of the page image. A text-to-page height ratio is determined and compared to one or more thresholds for determining legibility. In another aspect of the invention, a measure of the text in a page image is obtained by measuring the word density in the page image, while the measure derived from the page image comprises compressing the page image and determining the size of the compressed image file. Legibility is then determined by comparing the measure of word density with the compressed image file size.