X-Ray Image Contrast Alignment for Stable Region Identification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The accuracy of identifying regions in X-ray images using trained models is compromised due to discrepancies in imaging conditions between training data and actual image capture, caused by aging degradation or replacement of X-ray imaging apparatus, leading to inconsistent X-ray energy and dose.
Innovation Solution
An X-ray image preprocessing method that adjusts contrast by setting representative values of image regions to match training data conditions, using logarithmic transformation, region acquisition, and contrast adjustment to align with training data characteristics.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the same tube voltage and tube current are used for capturing training data and actual X-ray images, then the imaging conditions appear consistent, but aging degradation or replacement of the X-ray irradiation unit causes actual X-ray energy and dose to differ, reducing analysis accuracy
Solution Approach 1:
The patent applies parameter changes by adjusting the contrast of the X-ray image through logarithmic transformation and histogram equalization to compensate for variations in X-ray energy and dose caused by aging degradation or equipment replacement. This transforms the image characteristics to match the training data distribution, maintaining analysis accuracy despite changes in imaging parameters.
Solution Approach 2:
The patent implements feedback by using the trained model to evaluate the actual imaging conditions and adjusting the image processing parameters accordingly. The system compares the actual image characteristics with the training data characteristics and applies corrective transformations to align them, creating a closed-loop system that maintains consistency.
2Adaptability or versatility
If multiple trained models are generated based on training data captured under multiple imaging conditions, then coverage of different imaging conditions improves, but it becomes difficult to select the appropriate trained model for actual analysis
Solution Approach 1:
The patent applies universality by creating a single trained model that can handle multiple imaging conditions through contrast adjustment. Instead of generating multiple specialized models for different imaging conditions, the system uses one universal model combined with image processing transformations that adapt the image to match the training data characteristics, eliminating the need for model selection.
Solution Approach 2:
The patent introduces an intermediary image processing layer between the X-ray image capture and the trained model analysis. This intermediary applies contrast adjustment and logarithmic transformation to bridge the gap between actual imaging conditions and training data conditions, allowing a single model to work across different conditions without direct model selection.
Data Source
AI summary
An X-ray image preprocessing method is performed prior to a process of acquiring a region of an inspection object 90 from an X-ray image 40 using a trained model 31. The method comprises a step of acquiring the X-ray image 40, a step of acquiring a plurality of regions from the X-ray image 40, a step of acquiring an X-ray image representative value 50 of pixel values of each of the plurality of regions, and a step of acquiring a contrast-adjusted X-ray image 44 by performing contrast adjustment on the X-ray image 40 such that each X-ray image representative value 50 becomes a predetermined value 70 acquired based on input training data 32 used when creating the trained model 31.