White Balance Correction Using Pseudo-Neutral Zone Pixel Selection

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing white balance correction algorithms often incorrectly apply corrections to images, either overcorrecting or undercorrecting for white balance deviations, particularly in monochromatic images or when the average color is gray but local areas have tinted regions.

Innovation Solution

A method that selectively scans pixels in pseudo-neutral zones of an image, using intensity thresholds and neighborhood comparisons to determine relevant correction information, focusing on areas with consistent light intensity across different color channels to accurately assess and correct white balance deviations.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If conventional white balance correction algorithms are applied to all captured images, then white balance deviation can be corrected in images with gray average color, but incorrect corrections are applied to monochromatic images where no white balance deviation exists

Engineering Contradiction:

Improveaccuracy of white balance correctionVSAvoidincorrect white balance correction

Core Design Contradiction:

ReliabilityVSObject-generated harmful factors

Solution Approach 1:

The patent applies local quality by analyzing specific regions (pseudo-neutral zones) rather than the entire image. It identifies pixels with consistent intensity across color channels in localized areas to determine if white balance correction is needed, rather than making a global decision based on overall image statistics. This allows correct identification of monochromatic images as not requiring correction while still detecting local white balance deviations in mixed-content images.

Inventive Principle:

Principle #3Local quality

Solution Approach 2:

The patent segments the image analysis process into distinct steps: first identifying pixels with intensity above a threshold, then grouping them into pseudo-neutral zones based on spatial proximity and intensity consistency, and finally determining white balance correction needs based on these segmented regions. This segmentation allows the algorithm to focus computation on relevant image portions and avoid misclassifying monochromatic images.

Inventive Principle:

Principle #1Segmentation

2Ease of operation

If white balance correction is applied based on average image color, then processing is simple and fast, but corrections are missed in images where average color is gray but local areas have white balance deviation

Engineering Contradiction:

Improvesimplicity of correction algorithmVSAvoiddetection of white balance deviation

Core Design Contradiction:

Ease of operationVSMeasurement precision

Solution Approach 1:

The algorithm segments the image into potential pseudo-neutral zones by identifying pixels with consistent intensity across color channels and grouping spatially adjacent pixels. This segmentation enables local white balance analysis without requiring complex global optimization, maintaining computational simplicity while improving detection accuracy for localized color deviations.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

Instead of analyzing the entire image or using complex global optimization, the algorithm performs partial analysis by focusing on pixels meeting specific intensity and consistency criteria. This partial action approach maintains simplicity while sufficiently detecting local white balance issues that would be missed by average-based methods.

Inventive Principle:

Principle #16Partial or excessive action

3Loss of information

If all pixels are analyzed for white balance correction, then comprehensive correction information is obtained, but computational resources and processing time increase significantly

Engineering Contradiction:

Improvecompleteness of correction informationVSAvoidprocessing speed

Core Design Contradiction:

Loss of informationVSProductivity

Solution Approach 1:

The algorithm extracts only the necessary information for white balance correction by selecting pixels with intensity above a threshold and consistent across color channels, then groups them into pseudo-neutral zones. This extraction approach obtains sufficient correction information without processing every pixel, thereby maintaining completeness of relevant data while reducing computational burden.

Inventive Principle:

Principle #2Taking out (Extraction)

Solution Approach 2:

The algorithm performs partial analysis by processing only pixels meeting specific criteria (intensity threshold and color consistency) rather than all pixels. This partial action provides adequate correction information for most practical cases while significantly improving processing efficiency compared to exhaustive analysis.

Inventive Principle:

Principle #16Partial or excessive action

Data Source

PatentUS8498031B2Correction of white balance deviation in a captured image

Publication Date: 2013.07.30 STMICROELECTRONICS FRANCE

AI summary

A method is provided for correcting an image. At least some pixels in a captured image are scanned and there is selected each scanned pixel having at least one associated light intensity value greater than an intensity threshold value based on a comparison between the light intensity values for color indicators associated with the scanned pixel. Three correction information items are obtained by summing the light intensity values associated with the pixels selected by color indicator, and a correction of the white balance deviation affecting the captured image is determined based on the correction information items. A pixel is selected if both the difference between a mean light intensity value and a lowest light intensity value associated with the scanned pixel and the difference between a highest light intensity value and the mean light intensity value associated with the scanned pixel are less than a difference threshold value.