Color Correction Data Generation for Position-Independent Color Matching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing color correction techniques fail to account for changes in camera position and light source color tone, leading to deviations in image color characteristics during actual photography compared to color matching performed before photographing.
Innovation Solution
A color correction data generating apparatus and method that uses a learned model to infer images under prescribed conditions, generating color correction data to align color characteristics across multiple cameras and light sources, regardless of positional or tonal changes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of time
If color matching is performed before photographing using a light source and chart, then color correction data can be generated in advance, but color deviations occur when the camera position or light source color tone changes during actual photographing
Solution Approach 1:
The system performs preliminary color matching by photographing a color chart with multiple cameras under controlled lighting conditions before actual photographing. This preliminary action captures the color characteristics of each camera and establishes baseline correction data, allowing the system to prepare transformation models in advance while accounting for potential variations in camera position and light source color temperature.
Solution Approach 2:
The system measures and compensates for changes in lighting parameters (color temperature, intensity) and camera parameters (position, angle) by comparing actual photographing conditions against the preliminary color matching conditions. The color correction lookup tables are dynamically adjusted based on these parameter changes to maintain color consistency across different shooting scenarios.
2Ease of manufacture
If a color correction LUT is generated in advance based on controlled lighting conditions, then color processing is simplified, but color accuracy deteriorates when lighting conditions differ during actual photographing
Solution Approach 1:
The color correction system transitions from static pre-generated LUTs to dynamic adjustment mechanisms. The system continuously monitors actual lighting conditions (color temperature, intensity) and camera parameters during photographing, then dynamically adjusts the color correction LUTs to match the current conditions while maintaining the simplified lookup table approach for efficient real-time processing.
Solution Approach 2:
The system incorporates feedback mechanisms by comparing the actual photographing conditions (light source color temperature, camera position) against the preliminary color matching conditions. Based on this feedback, the system automatically adjusts the color correction parameters to compensate for deviations, ensuring color accuracy is maintained despite changes in environmental conditions.
3Reliability
If conventional color matching techniques are used, then color correction can be performed under controlled conditions, but the technique fails to adapt to changes in photographing position and light source color tone
Solution Approach 1:
The color correction system is designed to handle multiple scenarios and conditions universally. It processes color chart images from multiple cameras, adapts to different light source color temperatures, compensates for varying camera positions and angles, and works with different photographing conditions. This multi-functional capability allows the same system to maintain color consistency across diverse shooting environments without requiring separate correction procedures for each scenario.
Data Source
AI summary
A color correction data generating apparatus capable of performing color matching not depending on changes in the photographing position of the camera and the color tone of the light source is provided. The color correction data generating apparatus, which obtains a plurality of images obtained by photographing a predetermined subject with a plurality of image pickup apparatuses, includes at least one processor and/or circuit configured to function as following units, an output unit inputting an image of a predetermined subject region included in the obtained image into a learned model that machine learning has been already performed, and outputting an inferred image obtained by inferring the image of the predetermined subject region included in an image obtained when photographing the predetermined subject under a prescribed photographing condition, and a generating unit generating color correction data that matches color characteristics of a plurality of inferred images outputted by the output unit.