RGBW Image Upscale Unit Reducing Logic Circuit Size

Resolve Bottlenecks,
Find Innovative Solutions
Generate 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

VSEngineering 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

Engineering Contradiction:
Improveimage up-scaling qualityVSAvoidlogic circuit size
Core Design Contradiction:
Measurement precisionVSDevice complexity

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.

Inventive Principle:
Principle #1Segmentation

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.

Inventive Principle:
Principle #3Local quality

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

Engineering Contradiction:
Improveadditional pixel data accuracyVSAvoidnumber of line memories
Core Design Contradiction:
Measurement precisionVSQuantity of substance

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.

Inventive Principle:
Principle #1Segmentation

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.

Inventive Principle:
Principle #27Cheap short-living objects (Disposable)

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

Engineering Contradiction:
Improveinterpolation accuracyVSAvoidcomputational quantity
Core Design Contradiction:
Measurement precisionVSProductivity

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.

Inventive Principle:
Principle #35Parameter changes

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.

Inventive Principle:
Principle #16Partial or excessive action

Data Source

PatentEP3012830B1Image up-scale unit and method
Publication Date: 2020.07.08 LG DISPLAY CO LTD
  • EP3012830B1 patent drawingFigure 1~3
  • EP3012830B1 patent drawingFigure 4
  • EP3012830B1 patent drawingFigure 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.