A5016458559- Numerical Methods and Algorithms
- Advanced MRI Techniques and Applications
- Control Systems and Identification
- Sparse and Compressive Sensing Techniques
- Medical Imaging Techniques and Applications
- Fault Detection and Control Systems
- Advanced Vision and Imaging
- Aerospace and Aviation Technology
- Robotic Path Planning Algorithms
- GNSS positioning and interference
- Image and Signal Denoising Methods
- Probabilistic and Robust Engineering Design
- Model Reduction and Neural Networks
- Advanced Neuroimaging Techniques and Applications
- Neural Networks and Applications
- Fuzzy Logic and Control Systems
- Matrix Theory and Algorithms
- Geophysics and Gravity Measurements
- Advanced Numerical Analysis Techniques
- Guidance and Control Systems
- Advanced Image Processing Techniques
- Real-time simulation and control systems
- Biomimetic flight and propulsion mechanisms
- Adversarial Robustness in Machine Learning
- Inertial Sensor and Navigation
Philips (Netherlands)
2020-2021
Philips (Finland)
2014-2019
Delft University of Technology
2005-2012
Compressed sensing-sensitivity encoding is a promising MR imaging acceleration technique. This study compares the image quality of compressed accelerated with conventional sequences.Patients known, treated, or suspected brain tumors underwent 3D T1-echo-spoiled gradient echo T2-FLAIR sequences in addition to corresponding acquisition as part their clinical imaging. Two neuroradiologists blinded sequence and patient information independently evaluated both acquisitions. The were on 4- 5-point...
Adaptive intelligence aims at empowering machine learning techniques with the additional use of domain knowledge. In this work, we present application adaptive to accelerate MR acquisition. Starting from undersampled k-space data, an iterative learning-based reconstruction scheme inspired by compressed sensing theory is used reconstruct images. We developed a novel deep neural network refine and correct prior assumptions given training data. The was trained tested on knee MRI dataset 2019...
The problem of output optimization within a specified input space neural networks (NNs) with fixed weights is discussed in this paper. (highly) nonlinear when activation functions are used. This global encountered the reinforcement learning (RL) community. Interval analysis applied to guarantee that all solutions found any degree accuracy guaranteed bounds. major drawbacks interval analysis, i.e., dependency effect and high-computational load, both present for NN optimization. Taylor models...
The appearance variation cue captures the in texture a single image. Its use for obstacle avoidance is based on assumption that there less such when camera close to an obstacle. For videos of approaching frontal obstacles, it demonstrated combining with optic flow leads better performance than using either alone. In addition, successfully used control 16-g flapping-wing micro air vehicle DelFly II.
Through the years researchers have developed many different forms of spacecraft attitude controllers ranging from simple linear to highly nonlinear ones. For Nonlinear Dynamic Inversion controllers, tracking performance depends on model plant dynamics. In this paper we explore response a controlled satellite with liquid sloshing and apply neural networks create an adaptive NDI controller. Feedforward are used any unknown system The fuel motion is modeled using mechanical often in field...
A visual cue is introduced that exploits the appearance of a single image to estimate proximity an obstacle. In particular, variation captures in texture and / or color image, based on assumption there less such when camera close Random sampling applied order evaluate fast enough for use robotics. It demonstrated randomly sampled can be complementary optic flow obstacle detection; combining two cues leads better detection performance. sufficient computational efficiency cue's utilization...
Estimating multi-modal pilot model parameters from experiment or simulation data requires solving a global nonlinear optimization problem with many local minimums. Using the traditional parameter estimation techniques, finding optimum is dependent on initial estimate. In this paper performed by using theory of interval analysis, which deals intervals numbers instead crisp numbers. Interval analysis has been shown to be an excellent tool for and it can guarantee that minimum cost function...
NAVIGATION is a quarterly journal published by the Institute of Navigation. The publishes original, peer-reviewed articles on all aspects positioning, navigation, and timing. also selected technical notes survey articles, as well papers exceptional quality drawn from Institute's conference proceedings.
Adaptive intelligence aims at empowering machine learning techniques with the extensive use of domain knowledge. In this work, we present application adaptive to accelerate MR acquisition. Starting from undersampled k-space data, an iterative learning-based reconstruction scheme inspired by compressed sensing theory is used reconstruct images. We adopt deep neural networks refine and correct prior assumptions given training data. Our results show that approach performs better than...
*† ‡ § ¶ A new trajectory optimization algorithm for the Terminal Area Energy Management phase is presented based on interval analysis. Through a branch and bound strategy, analysis able to yield guaranteed rigorous bounds global minimum, i.e., best possible trajectory. It does so by using intervals instead of crisp numbers arithmetic arithmetic. Even numerical roundoff errors introduced computers are considered do not affect rigor solution. The steering commands vehicle optimized in order...
Spacecraft formation flying is the new trend in space missions. Increased flexibility, lower cost, enhanced capabilities and fail-safe concepts are main drivers of spacecraft formations. One important aspect operating formations trajectory planning. In this paper interval analysis applied to find set global optimal trajectories such that consumed fuel minimal and/or equalized over formation. Moreover, time which rotation can be performed. Interval implements arithmetic guarantee optimum...
ObjectiveThis study aimed to find the optimal acceleration factor achievable with CS-SENSE for a clinical ankle protocol while maintaining comparable image quality.MethodsWe explored CS-SENSE, an T2-weighted, PD-weighted TSE-Dixon (coronal, axial and sagittal) T2-mapping (sagittal) sequences, on 3 T MRI-scanner. This contained three steps: (1) phantom test, (2) pilot test healthy volunteers, (3) anatomical assessment cohort of volunteers quantitative analysis. images (acceleration factors...
Level set methods are used to determine the reachable for a given time domain.In aerospace industry level example flight envelopes.A new method is presented which uses interval analysis give guaranteed bounds on solution.Unlike current grid-based does not use grid and has lower computation complexity.The product of using analysis.Both an inner outer bound can be derived.The step accuracy easily controlled by user automatically adapted during processing.Initial tests show that provides...
Since multiple MRI contrasts of the same anatomy contain redundant information, one contrast can be used as a prior for guiding reconstruction an undersampled subsequent contrast. To this end, several learning-based guided methods have been proposed. However, two key challenges remain - (a) requirement large paired training datasets and (b) lack intuitive understanding model's internal representation utilization shared information. We propose modular two-stage approach reconstruction,...
Motivation: Scans within an MR exam share redundant information due to the same underlying structures. One contrast can hence be used guide reconstruction of another, thereby requiring less measurements. Goal(s): Multimodal guided reduce scanning times. Approach: Our method exploits AI-based content/style decomposition in iterative algorithm. We explored this concept via numerical simulation and subsequently validated it on vivo data. Results: Compared a conventional compressed sensing...
This paper provides improvements on the interval integer ambiguity resolution algorithm BOUNDS, which can give theoretical guarantees finding correct integers. The BOUNDS is validated by applying it to real GPS data and comparing LAMBDA method. first improvement a transformation of search space, reduces dependency. Secondly process using multiple frequencies explained demonstrated. third improvement, called contractions, quickly space increases solution accuracy. Finally, matrix computations...