Neural Network Illumination Map for Multi-Illuminant Color Cast Removal
Solution Overview
Problem
Existing image correction technologies struggle to effectively remove color casts caused by multiple illuminants in complex lighting environments, leading to suboptimal white balancing results.
Innovation Solution
A processor-implemented method using a neural network model to generate an illumination map that accounts for color casts from multiple illuminants, allowing for the generation of a white-adjusted image by removing portions of these color casts.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional white balancing methods are used, then the process is simple and fast, but the color cast removal accuracy is insufficient in multi-illumination environments
Solution Approach 1:
The illumination map is segmented into multiple partial illumination maps, each corresponding to a specific illuminant. This allows the system to process and correct color casts from different illuminants independently, improving accuracy while maintaining manageable complexity through modular processing
Solution Approach 2:
The illumination map serves as an intermediary representation that captures chromaticity information from multiple illuminants. This intermediate structure enables accurate color cast removal by providing detailed illumination characteristics without requiring direct complex processing of the original multi-illumination scene
2Reliability
If all color casts are removed completely, then white balancing is achieved, but chromaticity information from specific illuminants is lost
Solution Approach 1:
The system applies different correction strategies to different regions of the illumination map. For each illuminant, it can choose to remove color cast completely in some regions while preserving chromaticity information in other regions, allowing flexible control over information retention based on local image characteristics
Solution Approach 2:
Instead of completely removing all color casts, the system applies partial correction by removing only the portion of color cast that corresponds to unwanted illuminants while preserving the chromaticity information of desired illuminants. This partial action approach maintains necessary chromaticity data while achieving adequate white balancing
Data Source
AI summary
A method and apparatus with image correction is provided. A processor-implemented method includes generating, using a neural network model provided an input image, an illumination map including illumination values dependent on respective color casts by one or more illuminants individually affecting each pixel of the input image, and generating a white-adjusted image by removing at least a portion of the color casts from the input image using the generated illumination map.