Object Relighting Using Peripheral Luminance to Prevent Brightness Overflow
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing relighting processing techniques suffer from low accuracy in distance information, leading to unnatural images where brightened regions extend beyond the targeted object region, especially under challenging image capture conditions.
Innovation Solution
An image processing apparatus and method that controls virtual light source parameters based on object distance, luminance, and peripheral luminance to minimize unnaturalness by calculating virtual light reflection components and adjusting brightness levels.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If relighting processing is performed using distance information, then shadow correction and brightness adjustment are achieved, but the accuracy of distance information is low under challenging image capture conditions, leading to unnatural images where brightened regions extend beyond the targeted object region
Solution Approach 1:
The patent introduces peripheral luminance as an intermediary parameter to indirectly determine object boundaries when direct distance information is unreliable. By analyzing the luminance transition at object peripheries, the system can accurately identify object regions without depending solely on low-accuracy distance data, thus preventing brightened regions from extending beyond object boundaries
Solution Approach 2:
The patent changes the parameter used for object region determination from relying primarily on distance information to using peripheral luminance characteristics. By detecting luminance transitions and analyzing peripheral pixel values, the system adapts to challenging capture conditions where distance information is unreliable, maintaining image naturalness through alternative parameter-based object boundary detection
2Illumination intensity
If virtual light source parameters are controlled based on distance information, then brightness correction is applied to objects, but low accuracy distance information causes brightness corrections to extend beyond the intended object region
Solution Approach 1:
The patent applies local quality by using peripheral luminance to define precise local object boundaries before applying brightness correction. By analyzing luminance transitions at each peripheral region, the system ensures that virtual light parameters are applied only to appropriate local areas, preventing overflow into background regions while maintaining accurate brightness correction within object boundaries
Solution Approach 2:
The patent implements feedback by using detected peripheral luminance characteristics to adjust and refine the object region mask before applying relighting. The system continuously refines the boundary definition based on luminance transition feedback, ensuring that brightness corrections are precisely confined to object regions even when initial distance information is inaccurate
Data Source
Figure 1
Figure 2
Figure 3
AI summary
An image processing apparatus comprises a detection unit (113) configured to detect region information of an object in an image, an acquisition unit (105) configured to detect a predetermined object region that is a part of the object from the region information of the object, and to acquire luminance information of a periphery of the predetermined object region, and a relighting unit (114) configured to perform processing for correcting brightness of the object in the image by adding an effect of virtual light. The relighting unit corrects the brightness of the object including the predetermined object region in accordance with luminance of the periphery of the predetermined object region.