X-Ray Image Decomposition Using CT-Trained Region Separation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for decomposing medical images, such as X-ray images, struggle to accurately separate bones from soft tissues due to superimposition, requiring specialized imaging equipment and lacking effective machine learning techniques.
Innovation Solution
A learning method that utilizes a three-dimensional CT image to generate pseudo X-ray images, applying machine learning to iteratively refine the decomposition of anatomical regions through parameter updates based on comparisons between projected images, incorporating techniques like deep learning and region extraction to enhance accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If a simple X-ray image is used for imaging, then the imaging equipment is simple and easy to operate, but the anatomical structures (bones and soft tissues) are superimposed and cannot be easily recognized
Solution Approach 1:
The patent applies segmentation by decomposing the superimposed X-ray image into separate anatomical regions (bones and soft tissues) using machine learning. The model segments the image to isolate different tissue types, enabling clear recognition of anatomical structures without requiring complex multi-energy imaging equipment.
2Measurement precision
If dedicated imaging equipment with multiple energy bands is used to decompose bone and soft tissue, then the decomposition accuracy is improved, but the device complexity and cost increase
Solution Approach 1:
The patent replaces the mechanical/physical imaging system (multi-energy X-ray equipment) with a computational approach using machine learning. Instead of using complex hardware with multiple energy bands to achieve tissue decomposition, the system uses a trained neural network model that processes standard X-ray images to separate and identify different anatomical regions.
3Device complexity
If machine learning is used to decompose simple X-ray images, then the equipment requirement is reduced, but the decomposition accuracy into multiple anatomical regions is insufficient
Solution Approach 1:
The patent applies preliminary action by pre-training the machine learning model using labeled training data that includes ground truth anatomical segmentations. This preliminary training phase enables the model to learn the complex relationships between X-ray intensities and anatomical structures, so that when deployed, it can accurately decompose images into multiple anatomical regions without requiring complex imaging equipment.
Data Source
AI summary
An embodiment of the present invention provides a learning method, a learning apparatus, and a learning program for decomposing a two-dimensional X-ray image into a variety of regions, and an image processing method, an image processing apparatus, and an image processing program using a result of learning. A learning method according to an aspect of the present invention includes generating a masked CT image from a three-dimensional CT image, and generating a third pseudo X-ray image from the masked CT image. Then, a second pseudo X-ray image generated by projecting an original CT image is compared with the third pseudo X-ray image as a reference, and a parameter for use in generating the second pseudo X-ray image is updated based on a comparison result. A processor repeats the processing of each step until a predetermined condition is satisfied.