Motion Vector Reliability Adjustment for Bokeh Differences
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional motion vector detection techniques, such as template matching, face challenges in accurately determining the reliability of motion vectors when there are differences in bokeh or in-focus distances between images, leading to potential underestimation of reliability due to fixed threshold values.
Innovation Solution
An image processing apparatus and method that detects motion vectors between images and determines their reliability based on the correlation evaluation value, further considering the difference in bokeh amount and in-focus distances of the imaging optical system, dynamically adjusting threshold values to account for variations in bokeh.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If fixed threshold values are used to evaluate correlation between images, then the evaluation process is simple and fast, but the reliability determination becomes inaccurate when bokeh amounts differ between images
Solution Approach 1:
The patent applies dynamics by transitioning from fixed threshold values to dynamic threshold values that are adjusted based on the detected bokeh amount. The determination unit dynamically changes the threshold for correlation evaluation according to the bokeh characteristics of each image, allowing the system to adapt to varying optical conditions and maintain accurate reliability determination across different imaging scenarios.
Solution Approach 2:
The patent implements parameter changes by modifying the evaluation threshold parameter based on the bokeh amount parameter. When the bokeh amount is detected to be large, the threshold is adjusted accordingly, and when it is small, a different threshold is applied. This parameter adaptation enables the system to account for optical characteristics and improve reliability assessment accuracy without requiring complex manual intervention.
2Reliability
If correlation evaluation is performed without considering bokeh differences, then the processing is faster and simpler, but motion vector reliability is underestimated when in-focus distances differ
Solution Approach 1:
The patent applies preliminary action by detecting the bokeh amount before performing the correlation evaluation. The detection unit first analyzes the bokeh characteristics of the images, and then the determination unit uses this information to set appropriate threshold values for the correlation evaluation. This preliminary detection ensures that the reliability assessment is prepared in advance with the correct parameters, avoiding time-consuming adjustments during the main processing.
Solution Approach 2:
The patent implements feedback by using the detected bokeh amount to influence the correlation evaluation threshold. The system continuously monitors the bokeh characteristics and adjusts the evaluation criteria accordingly, creating a feedback loop that ensures accurate reliability determination. This feedback mechanism allows the system to adapt to different imaging conditions and maintain high reliability assessment accuracy.
3Reliability
If dynamic threshold adjustment based on bokeh is implemented, then reliability determination accuracy improves, but the computational complexity and processing time increase
Solution Approach 1:
The patent applies local quality by adjusting the evaluation threshold locally based on the specific bokeh characteristics detected in each image region. Rather than using a single global threshold for all images, the system determines appropriate thresholds for each image pair based on their individual bokeh amounts. This localized adaptation ensures high reliability determination accuracy for each specific imaging condition while avoiding unnecessary computational overhead for images with similar characteristics.
Data Source
AI summary
A disclosed image processing apparatus detects a motion vector between a first image and a second image based on the correlation between the first image and the second image. When the degree of reliability of the detected motion vector is determined based on an evaluation value regarding the correlation, a difference in amount of bokeh between the first image and the second image is considered to improve accuracy of the degree of reliability of the motion vector.