RGBW Image Upscale Unit Reducing Logic Circuit Size
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image up-scaling methods require a large computational quantity and significant logic circuit size due to the need for interpolation of adjacent pixel data signals, particularly in bi-cubic and Lanczos interpolation methods, which increases complexity and resource requirements.
Innovation Solution
The proposed solution involves dividing RGBW data signals into color difference and brightness components and performing different interpolation methods for each, utilizing an RGB interpolator and a W interpolator to generate additional pixels based on human color perception characteristics, thereby reducing computational complexity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If bi-cubic interpolation or Lanczos interpolation is used to generate additional pixel data signals, then image up-scaling quality is improved, but computational quantity and logic circuit size increase significantly
Solution Approach 1:
The patent divides the image data into multiple components (luminance Y and chrominance CbCr) and applies different interpolation methods to each component. The luminance component uses bi-linear interpolation while the chrominance component uses simple averaging, segmenting the processing to reduce overall complexity while maintaining quality where it matters most.
Solution Approach 2:
The patent applies different levels of interpolation complexity to different parts of the image data based on human visual sensitivity. High-quality bi-linear interpolation is applied only to the luminance component where quality is critical, while simpler methods are used for chrominance, creating local quality optimization that reduces computational burden.
2Measurement precision
If bi-cubic interpolation or Lanczos interpolation is performed on 16 adjacent pixel data signals repeatedly, then additional pixel data accuracy is improved, but the number of line memories required increases
Solution Approach 1:
The patent segments the interpolation process into separate luminance and chrominance processing paths, allowing each to use minimal memory resources appropriate to its requirements. The luminance path uses bi-linear interpolation with minimal line memory, while chrominance uses simple averaging requiring even less memory.
Solution Approach 2:
The patent replaces expensive, high-capacity line memory requirements with simpler, temporary storage solutions that only hold the minimum data needed for each interpolation calculation, effectively using disposable computational resources rather than permanent large-capacity memory.
3Measurement precision
If a sync function defined by third degree polynomials is used to determine weight values of adjacent pixels, then interpolation accuracy is improved, but computational quantity largely increases
Solution Approach 1:
The patent changes the complexity parameter of the interpolation function based on the data type being processed. For luminance, it uses bi-linear interpolation (second degree), and for chrominance, it uses simple averaging (first degree), adapting the mathematical complexity to the specific requirements of each component.
Solution Approach 2:
The patent applies full bi-linear interpolation only where necessary (luminance component) and uses simpler methods elsewhere (chrominance component), performing partial interpolation action that is sufficient for the application's needs without the excessive computational burden of applying high-order polynomials to all data.
Data Source
Figure 1~3
Figure 4
Figure 5
AI summary
An image up-scale unit is disclosed which up-scales an input image by adding a plurality of additional pixels. The image up-scale unit includes: a RGB interpolator configured to generate red (hereinafter, R), green (hereinafter, G) and blue (hereinafter, B) data signals for each of the additional pixels on the basis of the R, G and B data signals of an arbitrary pixel among pixels of the input image which are adjacent to the respective additional pixel; and a W interpolator configured to generate a W (hereinafter, W) data signal for each of the additional pixel on the basis of W data signals of the pixels of the input image which are adjacent to the respective additional pixel.