Spatial Domain Data Hiding for Robust Medical Image Authentication
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing lossless data hiding techniques face limitations in embedding large amounts of data without distorting the image, particularly in medical images, and often result in salt-and-pepper noise or reduced visual quality, while traditional digital signature techniques are fragile and fail to maintain authenticity under incidental distortion.
Innovation Solution
A method that divides image blocks into subsets, calculates difference values, and modifies pixel values to embed data without altering the original image significantly, using error correction coding to ensure data integrity and avoid noise, while also embedding authentication data to maintain image authenticity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If data is embedded in image using traditional lossless marking techniques, then data hiding capacity is limited, but image visual quality deteriorates with salt-and-pepper noise
Solution Approach 1:
The image is divided into multiple non-overlapping blocks, and each block is further divided into two subsets (even-positioned pixels and odd-positioned pixels). This segmentation allows the embedding process to operate locally on subsets rather than globally, reducing the propagation of noise artifacts and enabling higher data capacity without excessive salt-and-pepper noise.
Solution Approach 2:
The patent applies different embedding strategies to different blocks based on their local characteristics. By calculating block difference values and comparing them to thresholds, the method adapts the embedding intensity locally, modifying pixel values only when necessary and within acceptable distortion limits, thereby maintaining visual quality while maximizing data capacity.
2Quantity of substance
If pixel values are modified to embed data, then data hiding capacity increases, but image distortion increases
Solution Approach 1:
The patent modifies pixel values partially rather than completely. By dividing blocks into subsets and only modifying pixels in one subset based on block difference values, the method achieves data embedding with minimal distortion. The modification is applied selectively to achieve the necessary embedding capacity while staying within acceptable distortion thresholds.
Solution Approach 2:
The patent changes pixel value parameters selectively based on block characteristics. By calculating block difference values and comparing them to thresholds, the method determines whether and how much to modify pixel values in each block, adapting the parameter changes locally to balance embedding capacity with distortion control.
3Reliability
If lossless recovery is implemented, then original image can be restored, but data capacity is severely limited
Solution Approach 1:
The patent moves from traditional transform-domain methods to spatial domain operations, utilizing the block difference value dimension. By operating on subset pixel values and their differences rather than transform coefficients, the method creates additional embedding dimensions that enable both high data capacity and lossless recovery simultaneously.
Solution Approach 2:
The patent embeds data within the structure of the image by nesting the embedded data within block difference values. The embedded data is concealed within the natural variations of pixel differences, allowing high-capacity hiding while maintaining the ability to perfectly reconstruct the original image by reversing the subset modification process.
4Reliability
If authentication data is embedded to maintain authenticity, then robustness against distortion improves, but embedding complexity increases
Solution Approach 1:
The patent combines data hiding and authentication functions into a single embedding process. The same block difference value mechanism serves both to hide data and to provide authentication information, eliminating the need for separate authentication systems and reducing overall complexity while maintaining robustness against incidental distortion.
Data Source
AI summary
A method including identifying at least two subsets of pixels within a block of an image; forming a plurality of pixel groups from the at least two subsets of pixels, each pixel group having at least one pixel from a first of the at least two subsets and at least one pixel from a second of the at least two subsets; producing a plurality of difference values, each pixel group providing one of said difference values, each difference value being based on differences between pixel values of pixels within one of the pixel groups; and modifying pixel values of pixels in less than all of the at least two subsets, thereby embedding a bit value into the block.