Foreground Segmentation Using Dual Feature Thresholds
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing foreground segmentation methods using background subtraction, such as Japanese Patent Laid-Open No. 2013-186817, face challenges in accurately determining the foreground area due to threshold value limitations, leading to defects and enlargement of the foreground area, especially in cases with lens blur or aberrations, resulting in reduced image quality.
Innovation Solution
An image processing apparatus that calculates two feature amounts from input and background images to determine a threshold value for each pixel, using a two-dimensional table to adjust the threshold based on luminance differences and background luminance values, thereby improving the accuracy of foreground segmentation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If a simple threshold value is used for foreground determination, then the processing is simple, but the foreground area cannot be determined appropriately (defects and enlargement occur)
Solution Approach 1:
The patent changes the parameter used for threshold determination from a single feature amount (variance) to multiple feature amounts (variance and standard deviation). This allows the system to maintain simple processing while improving foreground determination accuracy by considering multiple characteristics of pixel values simultaneously.
Solution Approach 2:
The patent combines multiple feature amounts (variance and standard deviation) to form a composite threshold determination mechanism. This composite approach enables more accurate foreground area determination by integrating different statistical properties of pixel values, preventing both defects and enlargement in the segmented foreground area.
2Measurement precision
If the threshold value is reduced to capture blur area pixels, then more foreground pixels are detected, but blur area pixels are incorrectly determined as foreground (reduced accuracy)
Solution Approach 1:
The patent introduces a new parameter (standard deviation) in addition to the existing parameter (variance) for threshold determination. This parameter change enables the system to distinguish between actual foreground pixels and blur area pixels, allowing accurate detection of foreground pixels while maintaining high determination accuracy by using the combined information from both parameters.
3Device complexity
If a single feature amount (variance) is used for threshold calculation, then the processing is simple, but the foreground area cannot be determined appropriately in all cases
Solution Approach 1:
The patent creates a composite threshold calculation system that integrates multiple feature amounts (variance and standard deviation). This composite structure improves adaptability to different foreground scenarios while maintaining reasonable processing complexity, as the system can effectively handle various cases including those with lens blur and aberrations.
Solution Approach 2:
The patent makes the threshold calculation system universal by using multiple feature amounts that can adapt to different imaging conditions and foreground types. This multi-functional approach allows the same system to effectively segment foreground areas in diverse scenarios, including images with lens blur, aberrations, and varying lighting conditions.
Data Source
AI summary
An image processing apparatus generates an image indicating a foreground area of a captured image, based on (i) a threshold value corresponding to both luminance difference between the captured image and a background image and luminance of one of the captured image and the background image, (ii) obtained luminance of one of the captured image and the background image, and (iii) obtained luminance difference between the captured image and the background image.