Image Scaling Method Preserving Fine Details
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image scaling methods, such as linear interpolation and pixel sub-sampling, result in unwanted gray pixels and loss of fine details, especially when printing, leading to inferior image quality and non-uniform edges.
Innovation Solution
An edge-preserving image scaling method that considers additional features like edge contrast, rather than just average gray value, to preserve fine details and prevent the introduction of gray pixels, using a combination of hardware and programming to process images efficiently, allowing for real-time print operations without degrading image quality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If linear interpolation or pixel sub-sampling is used for image scaling, then the scaling process is simple and fast, but gray pixels are introduced and fine details are lost
Solution Approach 1:
The patent changes the scaling parameters by using a scaling factor that is not equal to an integer (e.g., 0.5), which requires handling non-integer pixel coordinates. This is resolved by using bicubic interpolation that calculates pixel values at non-integer positions, thereby avoiding the introduction of gray pixels while maintaining fine details.
Solution Approach 2:
The patent replaces simple interpolation or sub-sampling methods with bicubic interpolation, which uses a more sophisticated mathematical approach to calculate pixel values. This substitution eliminates the harmful effect of gray pixel introduction while preserving fine details and high-contrast features.
2Device complexity
If simple scaling algorithms are used, then computational overhead is low, but image quality degrades with loss of fine details and introduction of gray edges
Solution Approach 1:
The patent uses bicubic interpolation which involves calculating pixel values at non-integer coordinates using a 4x4 kernel. This changes the computational parameters from simple averaging to a more complex weighted sum operation, thereby improving image quality while managing computational overhead.
Solution Approach 2:
The patent introduces an intermediary calculation step where pixel values are computed at non-integer positions before being mapped to the output image grid. This intermediary step prevents direct sampling that would cause gray pixel introduction, thereby maintaining image quality.
3Manufacturing precision
If high-resolution scaling is performed, then image quality is maintained, but print speed decreases due to increased computational overhead
Solution Approach 1:
The patent performs scaling operations at the desired resolution using bicubic interpolation, which maintains image quality by accurately calculating pixel values. The method optimizes the process by handling the scaling in a single pass with efficient kernel calculations, thereby reducing computational overhead compared to multiple processing stages.
Data Source
AI summary
Examples include acquisition of an input image, generation of a first output pixel and a second output pixel based on a first set and a second set of input pixels, selective reassigning of a gray value of at least one of the first and second output pixels, and generation of an output image based on the output pixels at a second resolution lower than the first resolution.