Raster Image Edge Sharpness Preservation via Reparameterized Surface
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional image resampling methods, such as interpolation, often result in edge blurring during image magnification, degrading the visual quality of images by smoothing out sharp edges, which are crucial for maintaining image clarity.
Innovation Solution
A method that defines a parameterized surface interpolating the input raster image, where image edge locations are used to reparameterize the surface, preserving edge sharpness by introducing creases along detected edge curves, ensuring global C2 continuity while maintaining edge sharpness.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Volume of moving object
If conventional interpolation methods are used for image magnification, then the image can be resampled to a larger size, but edge blurring occurs and visual quality degrades
Solution Approach 1:
The patent applies different interpolation strategies to different regions of the image: standard interpolation in smooth regions and edge-preserving interpolation near detected edges. This local differentiation allows the system to maintain edge sharpness while still achieving image magnification, resolving the contradiction between increasing image size and preserving edge quality
Solution Approach 2:
The patent performs edge detection and classification before the actual resampling process. By identifying edge locations and orientations in advance, the system can prepare edge-preserving interpolation kernels specifically for edge regions, ensuring that edges maintain their sharpness during magnification while other regions are handled with standard interpolation
2Reliability
If smoothing filters are applied during resampling, then aliasing is reduced, but edge sharpness is lost
Solution Approach 1:
The patent applies smoothing filters selectively only in non-edge regions where aliasing is the primary concern, while using edge-preserving filters in regions containing edges. This localized filter selection allows the system to reduce aliasing in smooth areas without compromising edge sharpness, resolving the contradiction between reliability and manufacturing precision
Solution Approach 2:
The patent uses edge detection results as an intermediary to guide the selection and application of different filtering strategies. The edge map acts as a mediator that directs smoothing filters away from edge regions while allowing them to operate in smooth regions, thus reducing aliasing without sacrificing edge quality
Data Source
AI summary
Edges are detected in a raster image and generate parametric curves from the detected edges. The parametric curves are used to render a scaled version of the raster image. Some embodiments may allow edge locations within a raster image to retain a satisfactory level of sharpness when the raster image is scaled to a larger size.