Image Watermark Embedding Using Color Component Feature Indices
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional electronic watermark techniques face challenges in embedding and detecting codes in image data on portable devices due to high processing demands and adverse effects of unnecessary reflected wavelengths from CM components, leading to inaccurate code detection and image degradation.
Innovation Solution
An image data processing apparatus and method that divides data into blocks, extracts feature indices of color components, registers correspondence information, and embeds codes by changing the feature index of one color component based on another, adjusting pixel values to account for color characteristics and minimize the impact of unnecessary reflected wavelengths.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional electronic watermark technique using FFT is used to embed code into frequency region, then code embedding and detection can be achieved, but enormous amount of FFT calculation is required making it difficult to process within practical time on portable devices
Solution Approach 1:
The patent extracts only the necessary frequency information related to code embedding from the full FFT process. Instead of performing complete FFT calculations across all frequency components, the invention identifies and processes only specific frequency regions where code embedding occurs, significantly reducing computational load while maintaining detection accuracy.
Solution Approach 2:
The patent segments the frequency spectrum into relevant and irrelevant portions for code embedding. By dividing the frequency domain and focusing processing only on the segments containing embedded code information, the system avoids unnecessary FFT calculations on the entire frequency spectrum, thereby improving processing speed on portable devices.
2Measurement precision
If amplitude of pixel value fluctuation of Y component is set high to enable easy detection of embedded patterns, then code detection becomes easier, but the patterns become conspicuous on the printed image
Solution Approach 1:
The patent applies different amplitude levels of pixel value fluctuation in different spatial regions or frequency components of the image. By locally adjusting the fluctuation amplitude based on image characteristics and detection requirements, the system achieves sufficient code detection accuracy in critical regions while maintaining natural appearance in other areas, thus preventing conspicuous patterns.
Solution Approach 2:
The patent dynamically adjusts the amplitude parameter of pixel value fluctuation based on local image properties such as brightness, contrast, and texture. This parameter adaptation allows the embedded code to be detectable with sufficient amplitude while remaining imperceptible in regions where high amplitude would create visible artifacts.
3Measurement precision
If pixel value of Y component is adjusted based on densities of CMK components to compensate for unnecessary reflected wavelengths, then code detection accuracy improves, but the adjustment becomes complex due to printer driver modifications
Solution Approach 1:
The patent performs preliminary compensation for unnecessary reflected wavelengths from CM components during the code embedding stage. By pre-adjusting the Y component pixel values based on expected CMK density relationships before printing, the system eliminates the need for complex real-time adjustments during detection, simplifying the overall process while maintaining detection accuracy.
Solution Approach 2:
The patent creates a compensated model or lookup table that represents the relationship between CMK densities and Y component pixel values under ideal conditions. This pre-computed compensation model is then applied during code embedding and detection, avoiding the need for complex real-time calculations involving printer driver modifications.
Data Source
AI summary
A dividing unit divides image data into a plurality of blocks. An extracting unit extracts a feature index of a first color component and a feature index of a second color component in each of the blocks. A registration unit registers information about a correspondence between the feature index of a second color component and a change in the feature index for the first color component. A code embedding unit embeds a predetermined code into the image data, by changing the feature index of the first color component based on the feature index of the second color component, using the information registered.