A5059620666- Advanced X-ray and CT Imaging
- Medical Imaging Techniques and Applications
- Radiation Dose and Imaging
- Advanced X-ray Imaging Techniques
- Digital Radiography and Breast Imaging
- Radiomics and Machine Learning in Medical Imaging
- Advanced Radiotherapy Techniques
- Fault Detection and Control Systems
- Advanced MRI Techniques and Applications
- Radiology practices and education
- Advanced Statistical Process Monitoring
- Artificial Intelligence in Healthcare and Education
- Magnetic confinement fusion research
- Traumatic Brain Injury and Neurovascular Disturbances
- Biomedical and Engineering Education
- Atomic and Subatomic Physics Research
- Advanced Control Systems Optimization
- Laser-Plasma Interactions and Diagnostics
- Digital Holography and Microscopy
- Optical measurement and interference techniques
- Health Systems, Economic Evaluations, Quality of Life
United States Food and Drug Administration
2023-2024
Center for Devices and Radiological Health
2023-2024
Mayo Clinic in Arizona
2019-2024
Office of Science
2023-2024
Mayo Clinic
2017-2024
Food and Drug Administration
2023
WinnMed
2019-2022
Mayo Clinic in Florida
2020
Abstract In CT imaging of the head, multiple image series are routinely reconstructed with different kernels and slice thicknesses. Reviewing redundant information is an inefficient process for radiologists. We address this issue a convolutional neural network (CNN)-based technique, synthesiZed Improved Resolution Concurrent nOise reductioN (ZIRCON), that creates single, thin, low-noise combines favorable features from smooth sharp head kernels. ZIRCON uses CNN model autoencoder U-Net...
Abstract A key challenge for the development and deployment of artificial intelligence (AI) solutions in radiology is solving associated data limitations. Obtaining sufficient representative patient datasets with appropriate annotations may be burdensome due to high acquisition cost, safety limitations, privacy restrictions, or low disease prevalence rates. In silico offers a number potential advantages data, such as diminished harm, reduced simplified acquisition, scalability, improved...
Abstract Machine learning (ML) models often fail with data that deviates from their training distribution. This is a significant concern for ML-enabled devices as drift may lead to unexpected performance. work introduces new framework out of distribution (OOD) detection and monitoring combines ML geometric methods statistical process control (SPC). We investigated different design choices, including extracting feature representations quantification OOD in individual images an approach input...
We demonstrate a fast, flexible, and accurate paraxial wave propagation model to serve as forward for propagation-based X-ray phase contrast imaging (XPCI) in parallel-beam or cone-beam geometry. This incorporates geometric effects into the multi-slice beam method. It enables rapid prototyping is well suited tomographic reconstructions. Furthermore, it capable of modeling arbitrary objects, including those that are strongly multi-scattering. Simulation studies were conducted compare our...
In CT imaging, a standard set of bench testing performances, including modulation transfer function (MTF) and noise power spectrum (NPS), is considered essential to ensure that patient images from systems have sufficient image quality. These performances are measured using quality assurance (QA) phantoms commonly uniform background. However, for deep learning (DL)-based denoising models often overwhelmingly trained with images, it unclear whether background reflect the performance in...
Background: Machine learning (ML) methods often fail with data that deviates from their training distribution. This is a significant concern for ML-enabled devices in clinical settings, where drift may cause unexpected performance jeopardizes patient safety. Method: We propose Statistical Process Control (SPC) framework out-of-distribution (OOD) detection and monitoring. SPC advantageous as it visually statistically highlights deviations the expected To demonstrate utility of proposed...
A key challenge for the development and deployment of artificial intelligence (AI) solutions in radiology is solving associated data limitations. Obtaining sufficient representative patient datasets with appropriate annotations may be burdensome due to high acquisition cost, safety limitations, privacy restrictions or low disease prevalence rates. In silico offers a number potential advantages data, such as diminished harm, reduced simplified acquisition, scalability, improved quality...
Background SynthesiZed Improved Resolution and Concurrent nOise reductioN (ZIRCON) is a multi-kernel synthesis method that creates single series of thin-slice computed tomography (CT) images displaying low noise high spatial resolution, increasing reader efficiency minimizing partial volume averaging. Purpose To compare the diagnostic performance set ZIRCON to two routine clinical image using conventional CT head bone reconstruction kernels for diagnosing intracranial findings fractures in...
Purpose Demonstrate realistic simulation of grating‐based x‐ray phase‐contrast imaging (GB‐XPCI) using wave optics and the four‐dimensional Mouse Whole Body (MOBY) phantom defined with non‐uniform rational B‐splines (NURBS). Methods We use a full‐wave approach, which uses for propagation from source to detector. This forward model can be directly applied NURBS‐defined numerical phantoms such as MOBY. assign material properties (attenuation coefficient electron density) each part data adult...
Abstract Background Deep learning (DL) CT denoising models have the potential to improve image quality for lower radiation dose exams. These are generally trained with large quantities of adult patient data. However, CT, and increasingly DL methods, used in both pediatric populations. Pediatric body habitus size can differ significantly from adults vary dramatically newborns adolescents. Ensuring that subgroups different sizes not disadvantaged by methods requires evaluations capable...
X-ray dark-field measured on laboratory sources with large focal spots and detector apertures is sensitive to intra-pixel phase gradients abundant in the lungs due its hierarchical structure of subdividing airways terminating thin-walled alveoli. This work leverages this sensitivity exploit complementary information from x-ray attenuation computed tomography (CT) images improve quantification morphology pulmonary fibrosis. Specifically, a darkfield enhanced technique developed restore edges...
X-ray phase-contrast imaging (XPCI) overcomes the problem of low contrast between different soft tissues achieved in conventional x-ray by introducing phase as an additional mechanism. This work describes a compact light source (CXLS) and compares, via simulations, high quality XPCI results that can be produced from this to those using microfocus source. The simulation framework is first validated image acquired with microfocus-source, propagation-based (PB-XPCI) system. for water sphere...
We present a proof-of-principle demonstration of material decomposition using single X-ray tube potential (38 kVp + 0.2 mm Sn, for an effective energy around 27 keV) with Talbot-Lau grating-based phase-contrast computed tomography (CT) system. show good separation water and fat accurate quantitative measurement isopropyl alcohol. This method utilizes the distinctiveness both components refractive index, δ β, is promising separating soft tissue materials that have similar attenuation values...
Talbot-Lau grating interferometry enables the use of clinical x-ray tubes for phase contrast imaging, greatly broadening its utility both laboratory and preclinical applications. However, measurements made in porous or highly heterogeneous media are negatively impacted by low visibility, interferometer signal amplitude used to calculate relative shifts. While this loss visibility is source dark field it presents an additional noise images. In work, we develop a method normalized images as...