Neural Network Bone Density Estimation from Simple Radiation Images
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for estimating bone density, such as dual X-ray absorptiometry (DXA), are limited in accuracy and primarily provide two-dimensional assessments, which do not fully capture the complexity of three-dimensional bone density distribution.
Innovation Solution
A neural network-based estimation device and method that utilizes both simple radiation images and DXA scanning images, with energy subtraction techniques, to derive three-dimensional bone density by learning from teacher data including two-dimensional bone density information and imaging conditions, allowing for precise estimation of bone density across different energy distributions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If DXA method is used to measure bone density, then measurement can be performed with simple imaging, but the accuracy and three-dimensional information are insufficient
Solution Approach 1:
The patent transitions from two-dimensional bone density measurement to three-dimensional measurement by introducing multiple imaging angles and depths. The system captures images from different perspectives and combines them to reconstruct three-dimensional bone density distribution, thereby improving measurement accuracy without requiring overly complex specialized equipment.
Solution Approach 2:
The patent combines multiple imaging modalities and data sources (simple radiation images, DXA scanning images, images from different angles) into a unified three-dimensional bone density model. By merging these different types of imaging data, the system achieves high-accuracy three-dimensional measurement while utilizing existing imaging technologies.
2Measurement precision
If multiple imaging sources are used to improve accuracy, then estimation precision improves, but processing complexity increases
Solution Approach 1:
The patent introduces a trained neural network as an intermediary that automatically processes and integrates data from multiple imaging sources. The neural network learns the complex relationships between different imaging modalities and performs the integration task, reducing the need for manual processing and simplifying the overall system architecture while maintaining high accuracy.
Solution Approach 2:
The patent replaces traditional mechanical or manual data processing methods with a trained neural network. Instead of using complex algorithms or manual integration of multiple imaging sources, the system uses machine learning to automatically process and combine the data, thereby reducing processing complexity while improving accuracy.
3Ease of operation
If simple imaging is used, then ease of operation is maintained, but three-dimensional information is lost
Solution Approach 1:
The patent recovers three-dimensional information from two-dimensional simple imaging by introducing depth estimation through multiple angles and comparing images taken at different positions. This allows the system to maintain the simplicity of simple imaging operations while recovering the lost three-dimensional bone density information through computational methods.
Data Source
AI summary
A processor functions as a trained neural network that derives an estimation result relating to a three-dimensional 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 trained neural network learns using, as teacher data, (i) two radiation images or the like acquired by imaging the subject including the bone part with radiation having different energy distributions, and a two-dimensional bone density of the bone part included in the two radiation images or the like, or (ii) the radiation image or the like of the subject or a bone part image representing the bone part of the subject, the two-dimensional bone density of the bone part included in the radiation image or the like, or the bone part image, and the three-dimensional bone density of the bone part of the subject.