Neural Network Bone Density Estimation from Dual Imaging
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for estimating bone density, such as dual x-ray absorptiometry (DXA) and other radiation-based imaging techniques, face challenges in achieving high accuracy due to limitations in data resolution and the complexity of bone structure analysis.
Innovation Solution
An estimation device utilizing a learned neural network that processes both simple radiation images and DXA scanning images, combining three-dimensional CT images with derived attenuation coefficients to enhance bone density estimation, including bone density per unit area, fracture risk evaluation, and recovery state information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If simple radiation imaging is used for bone density estimation, then the imaging process is simple and fast, but the estimation accuracy is insufficient
Solution Approach 1:
The patent combines simple radiation images with DXA scanning images as dual input for the neural network. This merging of two different imaging modalities allows the system to maintain the speed advantage of simple imaging while incorporating the quantitative accuracy of DXA imaging, thereby resolving the contradiction between imaging speed and estimation accuracy.
Solution Approach 2:
The neural network serves as an intermediary that processes and integrates information from both simple radiation images and DXA scanning images. It learns to extract relevant features from both inputs and combines them to produce accurate bone density estimations, acting as a mediator that translates multiple image types into a unified accurate assessment.
2Measurement precision
If DXA scanning imaging is used for bone density estimation, then the estimation accuracy improves, but the imaging complexity increases
Solution Approach 1:
The system segments the imaging process into two parts: a simple radiation image for initial assessment and a DXA scanning image for detailed analysis. The neural network processes these segmented inputs separately and then integrates them, allowing the system to achieve high accuracy while maintaining operational simplicity through modular processing.
3Ease of manufacture
If neural network learning uses only simple radiation images, then the learning process is simple, but the estimation accuracy is limited
Solution Approach 1:
The training dataset is constructed as a composite of multiple data types: simple radiation images, DXA scanning images, and ground truth bone density values. This composite training material enables the neural network to learn from diverse inputs, improving estimation accuracy while the structured learning approach maintains relative simplicity in the learning process.
4Measurement precision
If multiple imaging modalities are combined for learning, then the estimation accuracy improves, but the data processing complexity increases
Solution Approach 1:
The neural network is designed with multi-functionality to handle both simple radiation images and DXA scanning images through a unified architecture. It performs multiple functions: processing different image types, extracting features from each modality, integrating the information, and producing bone density estimations. This universal design reduces processing complexity by using a single system for multiple tasks.
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
The solution enables more accurate bone density estimation and fracture risk assessment by leveraging composite two-dimensional images derived from CT data, improving diagnostic precision and recovery state monitoring.
Implementation Method 1
a learned neural network that derives a result of estimation relating to a bone density of a bone part from a simple radiation image acquired by simply imaging a subject including the bone part or a DXA scanning image acquired by imaging the subject by a DXA method
Implementation Method 2
a composite two-dimensional image representing the subject, which is derived by combining a three-dimensional CT image of the subject
Data Source
AI summary
An estimation device includes at least one processor, in which the processor functions as a learned neural network that derives a result of estimation relating to a bone density of a bone part from a simple radiation image acquired by simply imaging a subject including the bone part or a DXA scanning image acquired by imaging the subject by a DXA method. The learned neural network is learned by using, as teacher data, a composite two-dimensional image representing the subject, which is derived by combining a three-dimensional CT image of the subject, and information relating to the bone density of the subject.