Reversible Binary Image Data Hiding via Run-Length Histogram Modification

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing reversible binary image data hiding techniques face challenges in embedding data without distorting the original image, particularly with isolated white pixels that can damage reversibility and visual quality, and struggle to achieve high data embedding capacity while maintaining low changes in the marked image.

Innovation Solution

The use of run-length histogram modification combined with logical operations like AND and XOR to manipulate black and white run-lengths, allowing for reversible data hiding in binary images, which eliminates isolated white points and enhances embedding capacity, suitable for both non-halftone and halftone images.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Quantity of substance

If histogram modification is applied to embed data in binary images, then data embedding capacity is improved, but isolated white pixels are generated that damage reversibility

Engineering Contradiction:

Improvedata embedding capacityVSAvoidreversibility

Core Design Contradiction:

Quantity of substanceVSReliability

Solution Approach 1:

The patent applies preliminary action by performing row-division and word-division on the binary image before histogram modification. This preprocessing organizes the image into structured units (rows and words) that prevent isolated white pixels from being generated during data embedding, thereby maintaining reversibility while enabling data capacity expansion.

Inventive Principle:

Principle #10Preliminary action

2Quantity of substance

If more data is embedded in the binary image, then data embedding capacity is improved, but the visual quality and similarity to original image deteriorate

Engineering Contradiction:

Improvedata embedding capacityVSAvoidvisual quality

Core Design Contradiction:

Quantity of substanceVSManufacturing precision

Solution Approach 1:

The patent applies local quality by selectively modifying only specific regions of the binary image histogram rather than uniformly across the entire image. By targeting specific histogram bins and applying localized adjustments during data embedding, the method achieves high data capacity while minimizing visual distortion and maintaining image quality.

Inventive Principle:

Principle #3Local quality

Solution Approach 2:

The patent employs parameter changes by dynamically adjusting histogram modification parameters based on the embedding requirements and image characteristics. By changing parameters such as modification strength, target histogram bins, and division strategies, the system optimizes the balance between data embedding capacity and visual quality preservation.

Inventive Principle:

Principle #35Parameter changes

3Quantity of substance

If histogram modification is used for data hiding, then data embedding capability is improved, but the complexity of the processing method increases

Engineering Contradiction:

Improvedata embedding capabilityVSAvoidprocessing complexity

Core Design Contradiction:

Quantity of substanceVSDevice complexity

Solution Approach 1:

The patent applies segmentation by dividing the binary image into rows and words, and further segmenting the histogram into distinct bins and groups. This segmentation transforms the complex global histogram modification problem into multiple simpler local operations, making the processing more manageable and systematic while enhancing data embedding capability.

Inventive Principle:

Principle #1Segmentation

Data Source

PatentUS8270665B2System and method for reversible binary image data hiding using run-length histogram modification and logical operations

Publication Date: 2012.09.18 NEW JERSEY INSTITUTE OF TECHNOLOGY

AI summary

A method of hiding data comprising creating a histogram of an attribute of a first data set. The histogram includes occurrences of the attribute. Two adjacent occurrences are selected where a count of one of the two adjacent occurrences is zero. A second data set is embedded in data of the first data set associated with the selected adjacent occurrences.