A5069109153- Medical Imaging Techniques and Applications
- AI in cancer detection
- Digital Radiography and Breast Imaging
- Advanced X-ray and CT Imaging
- Advanced MRI Techniques and Applications
- Radiomics and Machine Learning in Medical Imaging
- Mathematical Dynamics and Fractals
- Advanced Topology and Set Theory
- Computability, Logic, AI Algorithms
- Advanced Neural Network Applications
- Advanced Radiotherapy Techniques
- Radiation Dose and Imaging
- semigroups and automata theory
- Medical Image Segmentation Techniques
- Particle Detector Development and Performance
- Cardiac Imaging and Diagnostics
- MRI in cancer diagnosis
- Adversarial Robustness in Machine Learning
- Domain Adaptation and Few-Shot Learning
- Nuclear Physics and Applications
- Sparse and Compressive Sensing Techniques
- Stochastic processes and statistical mechanics
- COVID-19 diagnosis using AI
- Cellular Automata and Applications
- Gamma-ray bursts and supernovae
The Netherlands Cancer Institute
2020-2025
Radboud University Nijmegen
2018-2024
Oncode Institute
2022-2024
Radboud University Medical Center
2019-2024
University of Amsterdam
2024
Delft University of Technology
2015-2017
Deep-learning-based brain magnetic resonance imaging (MRI) reconstruction methods have the potential to accelerate MRI acquisition process. Nevertheless, scientific community lacks appropriate benchmarks assess quality of high-resolution images, and evaluate how these proposed algorithms will behave in presence small, but expected data distribution shifts. The multi-coil (MC-MRI) challenge provides a benchmark that aims at addressing issues, using large dataset high-resolution,...
The two-dimensional nature of mammography makes estimation the overall breast density challenging, and true patient-specific radiation dose impossible. Digital tomosynthesis (DBT), a pseudo-3D technique, is now commonly used in cancer screening diagnostics. Still, severely limited 3rd dimension information DBT has not been used, until now, to estimate or dose. This study proposes reconstruction algorithm for based on deep learning specifically optimized these tasks. algorithm, which we name...
The goal of this work is to develop a data-driven empirical motion-artifact reduction algorithm for non-rigid motion in dedicated breast CT. Breast CT novel imaging modality that offers fully 3D images at good spatial resolution without compression and tissue overlap. However, the slow rotation speed gantry such systems increases likelihood artifacts. Because anatomy, motionartifact techniques need be able handle artifacts induced by motion, which cannot modeled due variable patterns...
Lesion volume is an important predictor for prognosis in breast cancer. However, it currently impossible to compute lesion volumes accurately from digital mammography data, which the most popular and readily available imaging modality We make a step towards more accurate measurement on mammograms by developing model that allows estimate processed mammogram. Processed are images routinely used radiologists clinical practice as well cancer screening medical centers. obtained raw mammograms,...
Abstract Background Computer algorithms that simulate lower‐doses computed tomography (CT) images from clinical‐dose are widely available. However, most operate in the projection domain and assume access to reconstruction method. Access commercial methods may often not be available medical research, making image‐domain noise simulation useful. introduction of non‐linear methods, such as iterative deep learning‐based reconstruction, makes insertion image intractable, it is possible determine...
Abstract Background Simulated computed tomography (CT) images allow for knowledge of the underlying ground truth and easy variation imaging conditions, making them ideal testing optimization new applications or algorithms. However, simulating all processes that affect CT can result in simulations are demanding terms processing time computer memory. Therefore, it is interest to determine how much simulation be simplified while still achieving realistic results. Purpose To develop a...
Digital breast tomosynthesis is rapidly replacing digital mammography as the basic x-ray technique for evaluation of breasts. However, sparse sampling and limited angular range gives rise to different artifacts, which manufacturers try solve in several ways. In this study we propose an extension Learned Primal- Dual algorithm tomosynthesis. The Primal-Dual a deep neural network consisting 'reconstruction blocks', take raw sinogram data initial input, perform forward backward pass by taking...
Cone beam computed tomography (CBCT) plays an important role in many medical fields nowadays. Unfortunately, the potential of this imaging modality is hampered by lower image quality compared to conventional CT, and producing accurate reconstructions remains challenging. A lot recent research has been directed towards reconstruction methods relying on deep learning, which have shown great promise for various modalities. However, practical application learning CBCT complicated several issues,...
Unpaired image-to-image translation has attracted significant interest due to the invention of CycleGAN, a method which utilizes combination adversarial and cycle consistency losses avoid need for paired data. It is known that CycleGAN problem might admit multiple solutions, our goal in this paper analyze space exact solutions give perturbation bounds approximate solutions. We show theoretically solution invariant with respect automorphisms underlying probability spaces, and, furthermore,...
We introduce computable actions of groups and prove the following versions effective Birkhoff’s ergodic theorem. Let Γ be a amenable group, then there always exists canonically tempered two-sided Følner sequence (F n ) n≥ 1 in Γ. For computable, measure-preserving, action on Cantor space $\{ 0,1\}^{\mathbb N}$ endowed with probability measure μ, it is shown that for every bounded lower semicomputable function f $\{0,1\}^{\mathbb {N}}$ Martin-Löf random $\omega \in \{0,1\}^{\mathbb equality...
It was shown by S. Kalikow and B. Weiss that, given a measure-preserving action of $\mathbb {Z}^d$ on probability space $\mathrm {X}$ nonnegative measurable function $f$ {X}$, the that sequence ergodic averages $$ \f
We investigate the use of deep learning in context X-ray polarization detection from astrophysical sources as will be observed by Imaging Polarimetry Explorer (IXPE), a future NASA selected space-based mission expected to operative 2021. In particular, we propose two models that can used estimate impact point well direction incoming radiation. The results obtained show data-driven approaches depict promising alternative existing analytical approaches. also discuss problems and challenges...
The purpose of this article is to extend the earliest results A.A. Brudno, connecting topological entropy a subshift X over $\mathbb{N}$ Kolmogorov complexity words in X, subshifts computable groups that posses Følner monotilings, which we introduce work. classical examples such are $\mathbb{Z}^d$ and upper-triangular matrices with integer entries. Following work B. Weiss show class closed under group extensions.
We re-examine the theory of systems with quasi-discrete spectrum initiated in 1960’s by Abramov, Hahn, and Parry. In first part, we give a simpler proof Hahn–Parry theorem stating that each minimal topological system
This thesis is dedicated to studying the theory of entropy and its relation Kolmogorov complexity. Originating in physics, notion was introduced mathematics by C. E. Shannon as a way measuring rate at which information coming from data source. There are, however, few different ways telling how much there is: an alternative approach quantifying amount complexity, proposed A. N. Kolmogorov. The key ingredient definition Kolmogorov-Sinai measure-preserving systems. Roughly speaking, expected...
Cone Beam CT (CBCT) is an essential imaging modality nowadays, but the image quality of CBCT still lags behind high standards established by conventional Computed Tomography. We propose LIRE+, a learned iterative scheme for fast and memory-efficient reconstruction, which substantially faster more parameter-efficient alternative to recently proposed LIRE method. LIRE+ rotationally-equivariant multiscale invertible primal-dual reconstruction. Memory usage optimized relying on simple reversible...
Cardiac MRI (CMRI) is a cornerstone imaging modality that provides in-depth insights into cardiac structure and function. Multi-contrast CMRI (MCCMRI), which acquires sequences with varying contrast weightings, significantly enhances diagnostic capabilities by capturing wide range of tissue characteristics. However, MCCMRI often constrained lengthy acquisition times susceptibility to motion artifacts. To mitigate these challenges, accelerated techniques use k-space undersampling via...