Image Processing Method for Signal-to-Noise Ratio and Resolution Trade-off

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing image sensors face a trade-off between increasing the signal-to-noise ratio and maintaining image resolution, as merging multiple photosensitive pixels reduces resolution while enhancing signal quality.

Innovation Solution

An image processing method that converts a color block image into an imitating image by determining whether pixel colors match, using direct pixel values for matching colors and interpolating from neighboring pixels for mismatched colors, while performing white balance compensation and restoration to enhance image quality.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If multiple photosensitive pixels are merged into a merged photosensitive pixel, then the signal-to-noise ratio of the merged image is increased, but the resolution of the merged image is decreased

Engineering Contradiction:
Improvesignal-to-noise ratioVSAvoidimage resolution
Core Design Contradiction:
ReliabilityVSManufacturing precision

Solution Approach 1:

The pixel array is divided into multiple pixel groups, where each pixel group includes multiple photosensitive pixels of different colors (e.g., R, G1, G2, B). By processing each pixel group separately and generating corresponding imitating pixels, the patent segments the image processing task to preserve resolution while maintaining the noise reduction benefits of pixel merging.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent applies different processing methods to different pixel locations based on local color characteristics. For pixels where the color matches the imitating pixel color, the original pixel value is used directly. For pixels where the color does not match, interpolation is performed using neighboring pixels of the same color. This local differentiation optimizes both resolution and signal-to-noise ratio.

Inventive Principle:
Principle #3Local quality

2Manufacturing precision

If pixel values are interpolated from neighboring pixels to maintain resolution, then the image resolution is maintained, but the signal-to-noise ratio may be compromised

Engineering Contradiction:
Improveimage resolutionVSAvoidsignal-to-noise ratio
Core Design Contradiction:
Manufacturing precisionVSReliability

Solution Approach 1:

The patent performs white balance compensation on the color block image before the interpolation process. This preliminary action ensures that the pixel values used for interpolation are already color-corrected, preventing color casts in the final image and maintaining both resolution and signal quality.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent performs white balance compensation restoration after interpolation to correct any color balance changes introduced by the interpolation process. This feedback mechanism ensures that the final imitating image maintains accurate color balance while benefiting from the resolution-preserving interpolation.

Inventive Principle:
Principle #23Feedback

Data Source

PatentEP3328080B1Image processing method and imaging apparatus
Publication Date: 2020.10.07 GUANGDONG OPPO MOBILE TELECOMMUNICATIONS CORP LTD
  • EP3328080B1 patent drawingFigure 1
  • EP3328080B1 patent drawingFigure 2
  • EP3328080B1 patent drawingFigure 3~4

AI summary

An image processing method includes determining whether a color of a current pixel of a first imitating image is same as a color of a related pixel of a color block image; when the same, using a pixel value of the related pixel as a pixel value of the current pixel; and when not the same, using pixel values of a related pixel unit to calculate the pixel value of the current pixel through interpolation. A color of the related pixel unit is same as the current pixel. A plurality of pixels of the related pixel unit are located in a first set of image pixel units of the color block image, wherein the first set of image pixel units neighbor a first image pixel unit that the related pixel belongs to. An imaging apparatus and a method performed by the imaging apparatus are also provided.