Image Scaling Using Virtual Pixel Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional image scaling methods, such as bicubic interpolation, face inefficiencies in computation and hardware implementation, leading to poor image magnification efficiency.
Innovation Solution
An image scaling method that groups adjacent pixels into non-intersecting sets, determines virtual interpolation pixels based on position relationships, calculates pixel sampling values using a bilinear interpolation algorithm, and generates theoretical pixel values to perform image scaling, effectively replacing bicubic interpolation with bilinear interpolation on virtual pixels.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If bicubic interpolation algorithm is used, then high-frequency information retention is improved, but computation complexity and hardware implementation difficulty increase
Solution Approach 1:
The patent segments the complex bicubic interpolation process into multiple bilinear interpolation steps. By dividing the 16 adjacent pixels into four groups and performing bilinear interpolation on each group separately, the computation is broken down into simpler, more manageable operations that reduce hardware complexity while maintaining interpolation accuracy.
Solution Approach 2:
The patent introduces virtual interpolation pixels as intermediary elements. Instead of directly computing bicubic interpolation from 16 adjacent pixels, the method first computes bilinear interpolation results to obtain virtual pixels, then uses these virtual pixels as intermediaries to achieve the final scaling effect. This intermediary approach simplifies the computational path.
2Measurement precision
If bicubic interpolation algorithm is used, then image magnification quality is improved, but image scaling efficiency deteriorates
Solution Approach 1:
The patent divides the image scaling process into segmented bilinear interpolation operations on grouped pixels rather than a single complex bicubic operation. This segmentation enables parallel processing of pixel groups, improving scaling efficiency while preserving image quality through the multi-step interpolation approach.
Solution Approach 2:
The patent creates virtual interpolation pixels as copies that represent the interpolated values. These virtual pixels are computed through bilinear interpolation and then used in subsequent processing steps, avoiding the need to repeatedly perform complex bicubic calculations and thereby improving processing efficiency.
3Measurement precision
If conventional bicubic interpolation is used, then interpolation accuracy is improved, but hardware implementation difficulty increases
Solution Approach 1:
The patent segments the interpolation computation into multiple simple bilinear steps that are easier to implement in hardware. By grouping adjacent pixels and applying bilinear interpolation sequentially, the hardware design becomes more straightforward compared to implementing a single complex bicubic interpolation unit.
Solution Approach 2:
The patent uses virtual interpolation pixels as intermediary data structures that simplify hardware implementation. These virtual pixels store intermediate bilinear interpolation results, allowing hardware to process data in simpler stages rather than requiring complex bicubic calculation units, thus easing manufacturing and implementation.
Data Source
AI summary
The present disclosure relates to an image scaling method and an apparatus thereof. The method includes: for at least one interpolation pixel in a source image, grouping at least two adjacent pixels of the interpolation pixel to obtain adjacent pixel sets; no intersection existing between the adjacent pixel sets; determining, based on position relationships between the interpolation pixel and regions formed by the adjacent pixel sets, virtual interpolation pixels corresponding to the interpolation pixel in the regions formed by the adjacent pixel sets respectively; calculating corresponding pixel sampling values of the virtual interpolation pixels in the source image according to a preset bilinear interpolation algorithm respectively; and generating theoretical pixel values corresponding to the interpolation pixel based on the pixel sampling values corresponding to the virtual interpolation pixels; the theoretical pixel values being configured to perform image scaling on the source image. The processing efficiency of the image scaling can be improved with the method.