Watermarking Dot Pattern for Image Visibility and Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image data generation techniques either result in poor visibility of the original image due to unintended patterns when embedding data or face difficulties in detecting embedded data with precision, especially when using linear dots with varying geometry and orientation.
Innovation Solution
The method involves generating synthesized image data by embedding data as a woven pattern with two-dimensionally arranged pixels in one direction and one-dimensionally arranged pixels in another direction, ensuring that the center positions of both types of dots are substantially equidistant, allowing for higher reproducibility and precise detection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Difficulty of detecting and measuring
If only one type of point dot is used to embed data, then data detection is facilitated, but unintended patterns are caused in the image resulting in poor visibility
Solution Approach 1:
The patent combines two types of dots (point dots and linear dots) to form a composite watermarking pattern. The point dots serve as reference markers while the linear dots encode data, creating a more robust system that avoids unintended patterns while facilitating detection through the structured relationship between the two dot types.
Solution Approach 2:
The watermarking pattern is segmented into two functional components: point dots that provide structural reference and linear dots that carry data information. This segmentation allows each component to perform its specific function optimally without interfering with the other, preventing unintended patterns while maintaining detectability.
2Object-affected harmful factors
If dots are arranged equidistantly to avoid unintended patterns, then image visibility is improved, but data detection becomes difficult requiring orientation determination processes
Solution Approach 1:
The patent introduces asymmetric linear dots with specific orientations to encode data, while maintaining equidistant arrangement of dot centers. The asymmetry in dot shape and orientation provides detectable features for data encoding without disrupting the regular spacing that prevents unintended patterns.
Solution Approach 2:
The patent transitions from two-dimensional point dots to one-dimensional linear dots, adding orientational information as an additional dimension for data encoding. This dimensional change enables data embedding through dot orientation while maintaining regular spatial spacing.
3Quantity of substance
If linear dots with varying geometry and orientation are used to embed data, then data capacity is increased, but reproduction stability and detection precision deteriorate
Solution Approach 1:
The patent applies different dot geometries locally: point dots are used at positions requiring high positional accuracy and as reference markers, while linear dots are used in positions where orientational data encoding is needed. This local differentiation optimizes both reproduction stability and detection precision for each dot type.
Data Source
AI summary
A PC generating image data divides data to be embedded into a predetermined size and arranges it in a predetermined area of a watermarking image. In the predetermined area are defined positions a-h assumed by an information dot configuring an information pattern based on the data to be embedded, and, of positions a-h, a position defined for the information pattern is assumed by a point dot. Of positions a-h, a position that is not defined for the information pattern is assumed by a linear dot. Furthermore in the predetermined area there is a position assumed by a positioning dot defining the predetermined area, and this position is assumed by a point dot. The linear dot has a geometry extending in one of the main scanning direction and vertical scanning direction of an image formation apparatus that allows higher performance in reproducibility for a line.