Radiation Image Processing With Multi-Energy Component Separation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing energy subtraction processing methods using two radiation images with different energy distributions struggle to accurately separate multiple components in a subject, such as fat and muscle, due to varying attenuation coefficients in soft tissues, limiting the separation of more than two components.
Innovation Solution
A radiation image processing device and method that utilizes n−1 types of radiation with different energy distributions to acquire and process n radiation images, deriving characteristics and thicknesses of components like soft parts and bone parts, enabling accurate separation of n components by enhancing component images based on body thickness and attenuation coefficients.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If energy subtraction processing uses two radiation images with different energy distributions, then two components (such as bone part and soft part) can be separated, but only two components can be obtained and cannot separate more components like fat and muscle
Solution Approach 1:
The patent transitions from two-dimensional energy subtraction (two images, two components) to three-dimensional decomposition (three images, three components). By introducing a third radiation image with a different energy distribution, the system adds a dimensional degree of freedom that enables separation of three components (bone, soft tissue, and either fat or muscle) through solving a system of three linear equations based on attenuation coefficients at three different energy levels.
2Ease of manufacture
If the attenuation coefficient of soft part is not obtained according to the ratio between fat and muscle, then processing is simplified, but accurate separation of multiple components becomes impossible
Solution Approach 1:
The patent implements an iterative feedback mechanism where the system initially estimates component ratios, calculates attenuation coefficients, performs decomposition, then uses the decomposition results to refine the ratio estimates. This feedback loop continues until convergence, allowing the system to automatically determine accurate fat-muscle ratios without manual input while achieving precise separation of all three components.
3Adaptability or versatility
If n radiation images with n different energy distributions are used, then n component images can be obtained, but the number of required radiation types increases
Solution Approach 1:
The patent makes a single radiation imaging system multi-functional by enabling it to capture images at three different energy distributions sequentially. The same imaging device performs multiple functions (acquiring low-energy, medium-energy, and high-energy images) by adjusting radiation parameters between shots, eliminating the need for multiple dedicated imaging systems while achieving three-component separation.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
Accurately separates n components in a radiation image by using n−1 types of radiation, enhancing component images to distinguish between soft parts, bone parts, and artificial objects, improving the precision of image separation beyond traditional two-component methods.
Implementation Method 1
The energy subtraction processing is a method in which respective pixels of the two radiation images obtained as described above are associated with each other, and subtraction is performed after multiplying a weight coefficient based on an attenuation coefficient in accordance with a component between pixels
Data Source
AI summary
A processor is configured to: acquire first to (n−1)th (n≥3) radiation images acquired by imaging a subject, which includes a first component consisting of a plurality of compositions and second to nth components each consisting of a single composition, with n−1 types of radiation having different energy distributions; derive a characteristic of the first component in at least a region of the subject in the first to (n−1)th radiation images; acquire a body thickness of the subject; derive thicknesses of the first to nth components by using the body thickness, the characteristic of the first component, and the first to (n−1)th radiation images; and derive first to nth component images in which the first to nth components are enhanced, respectively, based on the thicknesses of the first to nth components.