Pixel-Level White Balance Adjustment for Mixed Lighting
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current multi-area white balance processing techniques have limited flexibility in adjusting the white balance of images, especially when flash light and environmental light are involved, making it difficult to achieve precise color balance adjustments, particularly in night portrait scenes where users want to correct only specific portions of an image.
Innovation Solution
An image processing device that determines the types and influence rates of light sources in an image, acquires reference white balance gains for each light source, and calculates a mixture ratio to store these values with the image data, allowing for flexible white balance adjustments by applying the appropriate white balance gains to each pixel during post-processing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If a common white balance gain is applied to all pixels in an image, then the device complexity is reduced and processing is simplified, but the manufacturing precision of color balance correction deteriorates when multiple light sources are present
Solution Approach 1:
The patent divides the image into multiple regions based on light source characteristics, calculating separate white balance gains for each region. This segmentation allows different white balance corrections to be applied to areas affected by different light sources (e.g., flash light vs. environmental light), thereby improving color balance precision without requiring complete system redesign.
Solution Approach 2:
The patent applies local quality by determining white balance gains specific to each image region rather than using a uniform gain for the entire image. Each region's white balance gain is calculated based on its specific light source conditions, enabling precise local color correction while maintaining overall processing efficiency through automated region classification.
2Manufacturing precision
If multi-area white balance processing is applied to correct color balance in different image regions, then the manufacturing precision of color balance correction is improved, but the device complexity increases
Solution Approach 1:
The patent performs preliminary classification of image regions based on light source characteristics before calculating white balance gains. By pre-identifying which regions are affected by which light sources (flash vs. environmental light), the system prepares the necessary data structures and region masks in advance, reducing the computational complexity during the actual white balance calculation phase.
Solution Approach 2:
The system automatically detects and classifies light source types and affected regions without requiring manual intervention. The white balance processing unit self-determines the appropriate gains for each region based on embedded algorithms that analyze pixel values and light source characteristics, thereby reducing operational complexity while maintaining high precision.
3Adaptability or versatility
If white balance gains are calculated and stored for each light source type, then the adaptability of white balance adjustment is improved, but the loss of information increases due to additional data storage requirements
Solution Approach 1:
The patent creates a universal white balance processing framework that handles multiple light source types (flash, environmental light, mixed lighting) using a unified algorithmic approach. The same processing unit and data structures are used regardless of the specific lighting condition, allowing the system to adapt to various scenarios without requiring separate specialized systems for each light source type, thereby minimizing additional information storage requirements.
Data Source
AI summary
The number of light sources of original image data and the types of light source are determined (S11). A reference white balance gain which is set for each light source type of the original image data is acquired (S12). An influence rate of each light source type is acquired for each pixel of the original image data and a mixture ratio of the reference white balance gains is acquired on the basis of the influence rate (S13). The reference white balance gains and the mixture ratio are stored in a storage medium so as to be associated with the original image data (S14). It is possible to acquire the white balance gain for each pixel required for multi-area white balance processing for the original image data from the reference white balance gains and the mixture ratio stored in the storage medium subsequently.