Imaging Apparatus Subject Extraction Using Luminance and Texture Variation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing techniques for extracting a subject region from images fail to accurately distinguish the subject from the background in environments with rapid ambient light or background changes, such as moving vehicles or outdoors, due to false signals generated by these changes.
Innovation Solution
An imaging apparatus that alternately switches between two different illumination conditions and captures images under these conditions, calculating luminance and texture variations to accurately extract the subject region by distinguishing between changes caused by illumination, background, or subject movement.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If difference image method is used to extract subject region, then extraction speed is improved, but accuracy deteriorates in environments with rapid ambient light or background changes
Solution Approach 1:
The patent segments the image processing into two independent analysis paths: luminance variation analysis (for detecting subject regions through illumination changes) and texture variation analysis (for detecting background changes through texture patterns). By dividing the analysis into these separate segments, the system can independently evaluate each type of variation and combine their results to eliminate false signals, thereby maintaining high extraction speed while improving accuracy in dynamic environments.
Solution Approach 2:
The patent changes the analysis parameters from单一的 luminance difference to a dual-parameter system comprising both luminance variation (reflecting illumination changes and subject presence) and texture variation (reflecting background texture patterns). This parameter transformation allows the system to distinguish between genuine subject regions and false signals caused by ambient light or background changes, resolving the accuracy problem while preserving extraction speed.
2Reliability
If interpolated image generation is used to compensate for movement, then false signals from subject movement are reduced, but device complexity increases
Solution Approach 1:
The patent extracts and analyzes texture information separately from luminance information. By taking out the texture variation analysis as an independent component, the system can detect background changes and subject movements without requiring complex interpolated image generation. This extracted texture analysis works in parallel with luminance analysis, eliminating false signals while keeping the processing architecture relatively simple.
Solution Approach 2:
Instead of performing full interpolated image generation to compensate for all types of movement, the patent applies partial action by using texture variation analysis specifically targeted at detecting background changes and movement patterns. This partial analysis approach provides sufficient false signal elimination without the computational burden of complete image interpolation, reducing device complexity while maintaining reliability.
3Device complexity
If luminance variation alone is used for subject identification, then processing simplicity is maintained, but accuracy deteriorates due to susceptibility to ambient light changes
Solution Approach 1:
The patent merges two types of analysis results: luminance variation (which indicates subject presence through illumination changes) and texture variation (which indicates background stability through texture patterns). By combining these two independent analyses, the system maintains the simplicity of luminance-based detection while adding texture-based verification to eliminate false signals from ambient light changes, thus improving accuracy without significantly increasing processing complexity.
Solution Approach 2:
The patent introduces texture variation analysis as an intermediary verification mechanism. Rather than directly relying on luminance variation alone, the system uses texture analysis as an intermediate check to confirm whether luminance changes are due to genuine subject presence or ambient light fluctuations. This intermediary layer improves identification accuracy while keeping the overall processing approach relatively simple through modular design.
Data Source
AI summary
A luminance variation (dI) pertaining to each pixel is calculated (21) using a plurality of captured images obtained by image capturing under different illumination conditions, a texture variation (dF) pertaining to each pixel is calculated(22) using a plurality of captured images obtained by image capturing at different time points, and a subject region is extracted (23) based on the luminance variation (dI) and the texture variation (dF). A variation in the texture feature (F) pertaining to each pixel between the images is calculated as the texture variation (dF). The subject can be extracted with a high accuracy even when there are changes in the ambient light or the background.