Automatic White Balancing with Chromaticity Measure

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing automatic white-balancing (AWB) algorithms for digital cameras face challenges in accurately estimating illumination chromaticity with minimal reliance on camera module characterization (CMC) data and without high-level image processing, especially in devices with limited processing capability or under power constraints.

Innovation Solution

A computer-implemented method that estimates the white point of raw image data by normalizing color channel values, determining chromaticity weights based on disparities and distances in the sensor chromaticity space, and accumulating weighted image data to achieve robust illumination chromaticity measurement, independent of CMC data and high-resolution requirements.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If the gray-edge algorithm is used to estimate white point, then white point estimation accuracy is improved, but processing complexity and computational requirements increase significantly

Engineering Contradiction:
Improvewhite point estimation accuracyVSAvoidprocessing complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent extracts only the essential chromaticity information from image blocks without requiring full high-resolution image processing or complex edge detection algorithms. By working directly with down-sampled image blocks and their chromaticity coordinates, the method eliminates the need for computationally intensive edge detection while maintaining white point estimation accuracy.

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

The patent uses down-sampled image blocks as simplified, low-cost representations of the full image. These reduced-resolution blocks require minimal processing power and memory, making the algorithm suitable for mobile devices with limited resources while still providing accurate white point estimation through chromaticity analysis.

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

2Measurement precision

If high-resolution image data is used for edge information extraction, then white point estimation accuracy is improved, but memory usage and processing power requirements increase

Engineering Contradiction:
Improveedge information accuracyVSAvoidmemory usage
Core Design Contradiction:
Measurement precisionVSQuantity of substance

Solution Approach 1:

The patent divides the image into multiple down-sampled blocks and processes each block independently to determine chromaticity weights. This segmentation allows the algorithm to work with small, manageable data units that require minimal memory, while collectively providing comprehensive white point estimation across the entire image.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent transitions from spatial domain processing (requiring high-resolution edge detection) to chromaticity space processing. By converting image block data into chromaticity coordinates and analyzing color relationships in this transformed space, the method achieves accurate white point estimation without requiring high-resolution spatial information.

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

3Measurement precision

If camera module characterization data is used for white balancing, then color accuracy is improved, but susceptibility to manufacturing variations and computational cost increase

Engineering Contradiction:
Improvecolor accuracyVSAvoidsusceptibility to CMC errors
Core Design Contradiction:
Measurement precisionVSReliability

Solution Approach 1:

The patent enables the camera system to perform self-calibration by analyzing the chromaticity distribution within the captured image itself. Rather than relying on pre-stored camera module characterization data that may contain manufacturing variations, the algorithm derives white point information directly from the image content, making the system adaptive to its specific hardware instance.

Inventive Principle:
Principle #25Self-service

Solution Approach 2:

The patent changes the approach from using fixed camera module parameters to dynamically calculating chromaticity weights based on actual image block data. By computing weights from the ratio of color channel values in each block and comparing against expected achromatic values, the method adapts to actual sensor characteristics without requiring pre-characterization data.

Inventive Principle:
Principle #35Parameter changes

Data Source

PatentEP3138283B1Automatic white balancing with chromaticity measure of raw image data
Publication Date: 2019.07.17 INTEL CORP
  • EP3138283B1 patent drawingFigure 1~2
  • EP3138283B1 patent drawingFigure 3A~3B
  • EP3138283B1 patent drawingFigure 4A~4B

AI summary

An apparatus and methods for estimating a chromaticity of illumination from raw image data. In an embodiment, one or more image chromaticity weight is determined based on a distance between the raw image data in a sensor chromaticity space and a nearest point within a locus of sensor illumination chromaticities. In a further embodiment, one or more image chromaticity weight is determined based on a disparity among normalized color channel values. In certain embodiments, image chromaticity estimates are utilized to determine a white point estimate for the raw image data. In embodiments, an electronic device including a camera estimates the chromaticity value of raw image data captured by the camera as part of an AWB pipeline. The electronic device may further determine, for example as part of the AWB pipeline, a white point estimate based, at least in part, on the raw image data chromaticity value estimate(s).