X-ray CT System Using Deep Learning for Streak Artifact Mitigation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current spectral CT configurations face challenges such as pulse pileup in photon-counting detectors, low-energy noise in dual-layer detectors, high costs and cross-scatter effects in dual-source and dual-detector systems, and expensive ultra-high frequency generators in fast kVp-switching configurations.
Innovation Solution
The implementation of a deep learning-based X-ray computed tomography system that uses sparse kVp-switching and energy-integrating detectors to generate sparse view projection data, applying a two-channel deep learning artificial neural network to mitigate artifacts and perform image-domain material decomposition.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If photon-counting detectors are used for spectral CT, then spectral imaging capability is improved, but pulse pileup occurs reducing measurement precision
Solution Approach 1:
The detector is divided into multiple layers (first layer and second layer) with different scintillator materials optimized for different energy ranges. This segmentation allows the system to handle a broader energy spectrum while reducing pulse pileup by distributing photon detection across multiple specialized layers rather than a single detector.
Solution Approach 2:
The system changes the energy discrimination parameters by applying different energy thresholds to different detector layers. The first layer uses a first energy threshold and the second layer uses a second energy threshold, allowing the system to differentiate photon energies and reduce pileup effects through parameter-based photon classification.
2Measurement precision
If dual-layer detector with scintillators is used, then energy resolution is improved, but low-energy noise increases
Solution Approach 1:
Different regions of the detector (first layer and second layer) are assigned different scintillator materials with specific properties optimized for their respective functions. The first layer uses a scintillator material optimized for detecting lower energy photons while the second layer uses a different material for higher energy photons, creating local quality variations that reduce low-energy noise while maintaining energy resolution.
3Adaptability or versatility
If dual-source and dual-detector configuration is used, then spectral imaging capability is improved, but device complexity and cost increase
Solution Approach 1:
The patent merges the functions of dual-source and dual-detector systems into a single detector assembly with multiple layers. Instead of using two separate X-ray sources and two separate detectors, the invention combines energy discrimination and spectral imaging capabilities into one integrated detector system, reducing overall device complexity while maintaining spectral imaging functionality.
4Adaptability or versatility
If fast kVp-switching is used, then spectral imaging capability is improved, but cost of ultra-high frequency generator increases
Solution Approach 1:
The patent replaces the mechanical/electrical fast kVp-switching system with a detector-based energy discrimination system. Instead of rapidly switching the X-ray tube voltage to achieve spectral imaging, the invention uses a multi-layer detector with energy thresholds to differentiate photon energies, eliminating the need for expensive ultra-high frequency generators while maintaining spectral imaging capability.
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
This approach reduces the costs and complexities associated with existing spectral CT methods, improves image quality by mitigating streak artifacts, and efficiently decomposes images into material components, while using simpler hardware and avoiding the limitations of photon-counting detectors.
Implementation Method 1
A radiation source, such as an X-ray tube, irradiates the body from one side
Implementation Method 2
The attenuation of the radiation that has passed through the body is measured by processing electrical signals received from the detector
Implementation Method 3
The attenuation of the radiation that has passed through the body is measured by processing electrical signals received from the detector
Data Source
Figure 1
Figure 2A~2B
Figure 3A~3B
AI summary
According to one embodiment, an X-ray computed tomography (CT) system acquires a first image and a second image. The first image is based on a first projection dataset corresponding to a group of discrete first views and acquired with the emission of X-rays of a first energy. The second image is based on a second projection dataset corresponding to a group of discrete second views different from the first views and acquired with the emission of X-rays of a second energy different from the first energy. The X-ray CT system generates a third image corresponding to the streak artifact-mitigated first image and a fourth image corresponding to the streak artifact-mitigated second image, by inputting the first image and the second image to a learned model adapted to output the third image and the fourth image based on the first image and the second image.