Multi-Channel X-Ray AI for Abnormal Structure Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing AI-based computer-aided diagnosis (CAD) algorithms for medical images are limited by the lack of integration with multi-layer X-ray detectors, which can provide complementary information across different detection layers, leading to suboptimal accuracy in identifying suspicious areas.
Innovation Solution
A computer-implemented method utilizing a trained function based on machine learning, particularly deep learning, that processes data from multiple channels or layers of X-ray detectors, such as multi-slice or photon-counting detectors, to enhance the detection of abnormal structures by leveraging spectral and material-resolving information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If AI-based CAD algorithms are used with single-layer X-ray detectors, then the system complexity is low, but the detection accuracy and reliability are limited
Solution Approach 1:
The patent transitions from single-layer to multi-layer detector architecture, adding a spatial dimension (depth) to the detection system. Multiple detector layers capture X-ray information at different depths, providing complementary data that improves detection accuracy without requiring a complete system redesign
Solution Approach 2:
The detection system is segmented into multiple independent detector layers, each capable of capturing X-ray information. This segmentation allows parallel processing of data from different layers through the trained function, improving overall detection reliability while maintaining manageable system complexity
2Reliability
If multi-layer detectors are used to capture complementary information, then the detection reliability improves, but the data processing complexity increases
Solution Approach 1:
The patent merges information from multiple detector layers through a unified trained function (neural network). Instead of processing each layer separately with different algorithms, the complementary data from all layers is combined and processed together, improving detection reliability while streamlining the processing architecture
Solution Approach 2:
The trained function serves multiple purposes: it processes data from individual layers, integrates complementary information across layers, and performs the final detection decision. This multi-functional approach reduces the need for separate processing pipelines for each layer, managing data processing complexity
3Measurement precision
If spectral information from multiple energy levels is utilized, then the material differentiation capability improves, but the computational requirements increase
Solution Approach 1:
The trained function is pre-trained on comprehensive datasets that include spectral information from multiple energy levels. This preliminary training embeds the material differentiation capability into the model's weight parameters, allowing the system to perform sophisticated spectral analysis during inference without requiring intensive real-time computation
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
Improves the accuracy of AI-based CAD algorithms by simultaneously utilizing information from different detection layers or channels, enabling more reliable identification of suspicious structures in medical images.
Implementation Method 1
In direct-converting X-ray detectors, X-rays or photons can be converted into electrical pulses by a suitable converter material
Implementation Method 2
a phosphor screen is located in front of and another behind the detector circuit. The light generated in each of the fluorescent screens is detected by the same detector circuit
Data Source
Figure 1~2
Figure 3~4
Figure 5~6
AI summary
The invention relates to a computer-implemented method (10) for detecting an abnormal structure in an examination area in conjunction with an X-ray image from an X-ray system (20), comprising the following steps: - Receiving (12) input data, wherein the input data relates to an X-ray image data set having multiple data channels, - Applying (13) a trained function to the input data, wherein the trained function is based on a machine learning method, the trained function being applied to at least two data channels with respect to detecting the abnormal structure, and generating output data, - Providing (14) the output data, wherein the output data comprises an abnormal structure of the examination area.