Omnifocal Image Generation Using Definition-Based Weighting
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing methods for generating omnifocal images struggle with inconsistent luminance values at joints and the inclusion of unnecessary luminance values from images that should not be referenced, leading to unwanted foreign substances appearing in the final image.
Innovation Solution
An image processing method that calculates image reference values for each pixel by comparing definition among corresponding pixels in multiple captured images, and uses weighting based on distance and definition to smoothly change luminance values, preventing unnecessary values from appearing in the omnifocal image, with the option to use a Gaussian function for adjusting weighting coefficients.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Stability of the object's composition
If smoothing is applied to reference values to generate smooth omnifocal images, then luminance consistency at image joints is improved, but unnecessary luminance values from non-referenced images may appear in the omnifocal image
Solution Approach 1:
The patent divides the image into multiple focal planes and processes each plane separately. By segmenting the reference value calculation and application, it maintains luminance consistency within each focal plane while preventing contamination from non-referenced images through weighted combination based on definition values.
Solution Approach 2:
The patent changes the parameter of reference value from a single smoothed value to a multi-level weighted value based on definition. This parameter transformation allows simultaneous achievement of smooth transitions and prevention of unwanted luminance values by using definition-based weighting.
2Adaptability or versatility
If multiple captured images are combined to generate omnifocal images, then all cells in the culture solution can be focused, but luminance values become inconsistent at the joints of the images
Solution Approach 1:
The patent transforms the reference value parameter from a single smoothed value to a multi-level weighted value based on definition. This allows smooth transitions at joints while maintaining luminance consistency by weighting pixels according to their definition values across different focal planes.
Solution Approach 2:
The patent introduces definition as an intermediary parameter that mediates between multiple captured images. By using definition-based weighting, it achieves smooth luminance transitions at joints while preventing contamination from non-referenced images.
3Stability of the object's composition
If reference values are smoothed to prevent abrupt changes, then image transitions become smooth, but foreign substances from non-referenced images may appear in the omnifocal image
Solution Approach 1:
The patent changes the reference value from a single smoothed value to a definition-based weighted value. This parameter transformation enables smooth transitions through weighted averaging while eliminating foreign substances by assigning zero or low weights to pixels from non-referenced images based on their definition values.
Solution Approach 2:
The patent applies different weighting qualities to different pixels based on their definition values and spatial relationships. High-definition pixels receive higher weights while low-definition pixels (potential foreign substances) receive lower weights, achieving both smoothness and reliability.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Definition is calculated for each pixel in a plurality of captured images while changing a focal position. For each coordinates, an image reference value indicating the number of a to-be-referenced captured image is determined. Then, a luminance value is calculated for each coordinates on the basis of the image reference value and the definition. At this time, the luminance value in the captured image indicated by the image reference value of peripheral coordinates is reflected in the luminance value in the captured image indicated by the image reference value of each coordinates. The calculated luminance values are combined to generate an omnifocal image. Accordingly, a smooth omnifocal image is generated. The image reference values themselves are not changed. This prevents unnecessary luminance values of pixels existing in captured images that ought not to be referenced from appearing in the omnifocal image.