Image Enlargement Apparatus Edge Correction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image enlargement methods suffer from blurriness and unnatural edge generation during resolution conversion, particularly causing white colors to appear thicker due to excessive edge enhancement, leading to visual degradation and an unnatural appearance in images.
Innovation Solution
An image enlargement apparatus that includes an interpolation unit, a feature value analysis unit, a correction amount calculation unit, and a synthesis unit to generate a high-resolution image by adjusting edge components and re-correcting pixel values based on gradient strengths and thresholds, thereby suppressing overshoot and undershoot while maintaining image sharpness.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If edge enhancement is performed using a high-pass enhancement filter, then image sharpness is improved, but overshoot and undershoot artifacts occur near edges causing visual degradation
Solution Approach 1:
The patent applies different processing strategies to different regions of the image based on local characteristics. Edge portions undergo enhancement processing to improve sharpness, while non-edge portions use standard interpolation to avoid artifacts. This localized approach ensures that edge sharpness is improved without introducing overshoot and undershoot in other regions.
Solution Approach 2:
The patent performs edge enhancement only partially - specifically only in edge portions where it is needed, rather than applying enhancement to the entire image. This partial action approach maintains image sharpness at edges while avoiding the harmful overshoot and undershoot artifacts that would result from excessive enhancement across the whole image.
2Object-generated harmful factors
If standard interpolation methods (bilinear or bicubic) are used, then block artifacts are suppressed, but the resulting image becomes blurred
Solution Approach 1:
The patent segments the image into edge portions and non-edge portions, then applies different interpolation methods to each segment. Edge portions receive enhancement processing to maintain sharpness, while non-edge portions use standard interpolation to avoid block artifacts. This segmentation allows the system to achieve both goals - suppressing blocks while maintaining sharpness in critical regions.
Solution Approach 2:
Different interpolation qualities are applied to different parts of the image. Edge regions receive higher-quality enhancement processing to preserve sharpness, while non-edge regions use standard interpolation methods. This local quality differentiation ensures that sharpness is maintained where it matters most (at edges) without introducing block artifacts in smooth regions.
3Device complexity
If nearest neighbor interpolation is used, then computational complexity is reduced, but visual blocks become noticeable causing degradation in image quality
Solution Approach 1:
The patent applies the computationally intensive enhancement processing only partially - specifically only to edge portions where it is most needed for quality, rather than to the entire image. This partial application reduces the overall computational burden compared to full-image enhancement while still eliminating visual blocks in critical edge regions.
Data Source
AI summary
An image enlargement apparatus includes an interpolation unit that generates a high-resolution interpolation image from a low-resolution image; and a synthesis unit that generates a first high-resolution image by synthesizing the interpolation image and a correction amount. Additionally, a re-correction unit generates a second high-resolution image by re-correcting the first high-resolution image using a generated feature value.