-
Flexible metasurface for improving brain imaging at 7T
Authors:
Vladislav Koloskov,
Wyger Brink,
Andrew Webb,
Alena Shchelokova
Abstract:
Ultra-high field MRI offers unprecedented detail for non-invasive visualization of the human brain. However, brain imaging is challenging at 7T due to the B$_1^+$ field inhomogeneity, which results in signal intensity drops in temporal lobes and a bright region in the brain center. This study aims to evaluate using a metasurface to improve brain imaging at 7T and simplify the investigative workflo…
▽ More
Ultra-high field MRI offers unprecedented detail for non-invasive visualization of the human brain. However, brain imaging is challenging at 7T due to the B$_1^+$ field inhomogeneity, which results in signal intensity drops in temporal lobes and a bright region in the brain center. This study aims to evaluate using a metasurface to improve brain imaging at 7T and simplify the investigative workflow. Two flexible metasurfaces, each comprising a periodic structure of copper strips and parallel-plate capacitive elements printed on an ultra-thin substrate, were optimized for brain imaging and implemented via PCB. We considered two setups: (1) two metasurfaces located near the temporal lobes; and (2) one metasurface placed near the occipital lobe The effect of metasurface placement on the transmit efficiency and specific absorption rate was evaluated via electromagnetic simulation studies with voxelized models. In addition, their impact on SNR and diagnostic image quality was evaluated in vivo for male and female volunteers. Placement of metasurfaces near the regions of interest led to an increase in homogeneity of the transmit field by 5$\%$ and 10.5$\%$ in the right temporal lobe and occipital lobe for a male subject, respectively. SAR values changed insignificantly and were under recommended limits. In vivo studies also confirmed the numerically predicted improvement in field distribution and receive sensitivity in the desired ROI. Optimized metasurfaces enable homogenizing transmit field distribution in the brain at 7T. The proposed lightweight and flexible structure has the potential to provide MR examination with higher diagnostic value images.
△ Less
Submitted 30 October, 2023;
originally announced October 2023.
-
An anthropomorphic thyroid phantom for ultrasound-guided radiofrequency ablation of nodules
Authors:
Tim Boers,
Wyger Brink,
Leonardo Bianchi,
Paola Saccomandi,
Johan van Hespen,
Germen Wennemars,
Sicco Braak,
Michel Versluis,
Srirang Manohar
Abstract:
Background: Needle-based procedures such as fine needle aspiration (FNA) and thermal ablation, are often applied for thyroid nodule diagnosis and therapeutic purposes, respectively. With blood vessels and nerves nearby, these procedures can pose risks in damaging surrounding critical structures. Purpose: The development and validation of innovative strategies to manage these risks require a test o…
▽ More
Background: Needle-based procedures such as fine needle aspiration (FNA) and thermal ablation, are often applied for thyroid nodule diagnosis and therapeutic purposes, respectively. With blood vessels and nerves nearby, these procedures can pose risks in damaging surrounding critical structures. Purpose: The development and validation of innovative strategies to manage these risks require a test object with well-characterized physical properties. For this work, we focus on the application of ultrasound-guided thermal radio-frequency ablation (RFA). Methods: We have developed an anthropomorphic phantom mimicking the thyroid and surrounding anatomical and physiological structures that are relevant to ultrasound-guided thermal ablation. The phantom was composed of a mixture of polyacrylamide, water, and egg white extract and was cast using molds in multiple steps. The thermal, acoustical, and electrical characteristics were experimentally validated. The ablation zones were analyzed via non-destructive T2-weighted MRI scans utilizing the relaxometry changes of coagulated egg albumen, and the temperature distribution was monitored using an array of fiber Bragg sensors. Results: The physical properties of the phantom were verified both on ultrasound as well as its response to thermal ablation. The final temperature achieved (92°C), the median percentage of the nodule ablated (82.1%), the median volume ablated outside the nodule (0.8 mL), and the median number of critical structures affected (0) were quantified. Conclusion: An anthropomorphic phantom that can provide a realistic model for development and training in ultrasound-guided needle-based thermal interventions for thyroid nodules has been presented. In the future, this model can also be extended to novel needle-based diagnostic procedures.
△ Less
Submitted 21 July, 2023;
originally announced July 2023.
-
Preoperative brain tumor imaging: models and software for segmentation and standardized reporting
Authors:
D. Bouget,
A. Pedersen,
A. S. Jakola,
V. Kavouridis,
K. E. Emblem,
R. S. Eijgelaar,
I. Kommers,
H. Ardon,
F. Barkhof,
L. Bello,
M. S. Berger,
M. C. Nibali,
J. Furtner,
S. Hervey-Jumper,
A. J. S. Idema,
B. Kiesel,
A. Kloet,
E. Mandonnet,
D. M. J. Müller,
P. A. Robe,
M. Rossi,
T. Sciortino,
W. Van den Brink,
M. Wagemakers,
G. Widhalm
, et al. (5 additional authors not shown)
Abstract:
For patients suffering from brain tumor, prognosis estimation and treatment decisions are made by a multidisciplinary team based on a set of preoperative MR scans. Currently, the lack of standardized and automatic methods for tumor detection and generation of clinical reports represents a major hurdle. In this study, we investigate glioblastomas, lower grade gliomas, meningiomas, and metastases, t…
▽ More
For patients suffering from brain tumor, prognosis estimation and treatment decisions are made by a multidisciplinary team based on a set of preoperative MR scans. Currently, the lack of standardized and automatic methods for tumor detection and generation of clinical reports represents a major hurdle. In this study, we investigate glioblastomas, lower grade gliomas, meningiomas, and metastases, through four cohorts of up to 4000 patients. Tumor segmentation models were trained using the AGU-Net architecture with different preprocessing steps and protocols. Segmentation performances were assessed in-depth using a wide-range of voxel and patient-wise metrics covering volume, distance, and probabilistic aspects. Finally, two software solutions have been developed, enabling an easy use of the trained models and standardized generation of clinical reports: Raidionics and Raidionics-Slicer. Segmentation performances were quite homogeneous across the four different brain tumor types, with an average true positive Dice ranging between 80% and 90%, patient-wise recall between 88% and 98%, and patient-wise precision around 95%. With our Raidionics software, running on a desktop computer with CPU support, tumor segmentation can be performed in 16 to 54 seconds depending on the dimensions of the MRI volume. For the generation of a standardized clinical report, including the tumor segmentation and features computation, 5 to 15 minutes are necessary. All trained models have been made open-access together with the source code for both software solutions and validation metrics computation. In the future, an automatic classification of the brain tumor type would be necessary to replace manual user input. Finally, the inclusion of post-operative segmentation in both software solutions will be key for generating complete post-operative standardized clinical reports.
△ Less
Submitted 29 April, 2022;
originally announced April 2022.
-
A Probabilistic Graphical Model Approach to the Structure-and-Motion Problem
Authors:
Simon Streicher,
Willie Brink,
Johan du Preez
Abstract:
We present a means of formulating and solving the well known structure-and-motion problem in computer vision with probabilistic graphical models. We model the unknown camera poses and 3D feature coordinates as well as the observed 2D projections as Gaussian random variables, using sigma point parameterizations to effectively linearize the nonlinear relationships between these variables. Those vari…
▽ More
We present a means of formulating and solving the well known structure-and-motion problem in computer vision with probabilistic graphical models. We model the unknown camera poses and 3D feature coordinates as well as the observed 2D projections as Gaussian random variables, using sigma point parameterizations to effectively linearize the nonlinear relationships between these variables. Those variables involved in every projection are grouped into a cluster, and we connect the clusters in a cluster graph. Loopy belief propagation is performed over this graph, in an iterative re-initialization and estimation procedure, and we find that our approach shows promise in both simulation and on real-world data. The PGM is easily extendable to include additional parameters or constraints.
△ Less
Submitted 7 October, 2021;
originally announced October 2021.
-
Towards the Localisation of Lesions in Diabetic Retinopathy
Authors:
Samuel Ofosu Mensah,
Bubacarr Bah,
Willie Brink
Abstract:
Convolutional Neural Networks (CNNs) have successfully been used to classify diabetic retinopathy (DR) fundus images in recent times. However, deeper representations in CNNs may capture higher-level semantics at the expense of spatial resolution. To make predictions usable for ophthalmologists, we use a post-attention technique called Gradient-weighted Class Activation Map** (Grad-CAM) on the pe…
▽ More
Convolutional Neural Networks (CNNs) have successfully been used to classify diabetic retinopathy (DR) fundus images in recent times. However, deeper representations in CNNs may capture higher-level semantics at the expense of spatial resolution. To make predictions usable for ophthalmologists, we use a post-attention technique called Gradient-weighted Class Activation Map** (Grad-CAM) on the penultimate layer of deep learning models to produce coarse localisation maps on DR fundus images. This is to help identify discriminative regions in the images, consequently providing evidence for ophthalmologists to make a diagnosis and potentially save lives by early diagnosis. Specifically, this study uses pre-trained weights from four state-of-the-art deep learning models to produce and compare localisation maps of DR fundus images. The models used include VGG16, ResNet50, InceptionV3, and InceptionResNetV2. We find that InceptionV3 achieves the best performance with a test classification accuracy of 96.07%, and localise lesions better and faster than the other models.
△ Less
Submitted 2 February, 2021; v1 submitted 21 December, 2020;
originally announced December 2020.
-
Mode I fracture toughness of asymmetric metal-composite adhesive joints
Authors:
Theodoros Loutas,
Panayiotis Tsokanas,
Vassilis Kostopoulos,
Peter Nijhuis,
Wouter M. van den Brink
Abstract:
In this work, the mode I fracture toughness of dissimilar metal-composite adhesive joints is experimentally investigated using the double cantilever beam (DCB) test. The particular joint under study is resulted by the adhesive joining of a thin titanium sheet with a thin carbon fiber reinforced plastic (CFRP) laminate and is envisioned to be implemented in the hybrid laminar flow control system of…
▽ More
In this work, the mode I fracture toughness of dissimilar metal-composite adhesive joints is experimentally investigated using the double cantilever beam (DCB) test. The particular joint under study is resulted by the adhesive joining of a thin titanium sheet with a thin carbon fiber reinforced plastic (CFRP) laminate and is envisioned to be implemented in the hybrid laminar flow control system of future aircraft. Four different industrial technologies for the joining of the titanium and CFRP adherents are evaluated/compared; co-bonding with and without adhesive and secondary bonding using either thermoset or thermoplastic CFRP. The vacuum-assisted resin transfer molding (VARTM) technique is employed for the manufacturing of the panels. After manufacturing, the panels are cut into test specimens that, because they are too thin (approximately 2.4 mm thick), needed to be stiffened from both titanium and composite sides with two aluminum backing beams to ensure the non-yielding of the titanium during the subsequent DCB tests. Towards the determination of the fracture toughness of the joint from the experimental data, an analytical model recently developed by the authors, that considers the bending-extension coupling of both sub-laminates constituting the test specimen as well as the manufacturing-induced residual thermal stresses, is applied. For the four manufacturing options (MO) investigated, the load-displacement behaviors, failure patterns, and fracture toughness performances are presented and compared.
△ Less
Submitted 7 May, 2020; v1 submitted 24 February, 2020;
originally announced February 2020.
-
BINet: a binary inpainting network for deep patch-based image compression
Authors:
André Nortje,
Willie Brink,
Herman A. Engelbrecht,
Herman Kamper
Abstract:
Recent deep learning models outperform standard lossy image compression codecs. However, applying these models on a patch-by-patch basis requires that each image patch be encoded and decoded independently. The influence from adjacent patches is therefore lost, leading to block artefacts at low bitrates. We propose the Binary Inpainting Network (BINet), an autoencoder framework which incorporates b…
▽ More
Recent deep learning models outperform standard lossy image compression codecs. However, applying these models on a patch-by-patch basis requires that each image patch be encoded and decoded independently. The influence from adjacent patches is therefore lost, leading to block artefacts at low bitrates. We propose the Binary Inpainting Network (BINet), an autoencoder framework which incorporates binary inpainting to reinstate interdependencies between adjacent patches, for improved patch-based compression of still images. When decoding a patch, BINet additionally uses the binarised encodings from surrounding patches to guide its reconstruction. In contrast to sequential inpainting methods where patches are decoded based on previons reconstructions, BINet operates directly on the binary codes of surrounding patches without access to the original or reconstructed image data. Encoding and decoding can therefore be performed in parallel. We demonstrate that BINet improves the compression quality of a competitive deep image codec across a range of compression levels.
△ Less
Submitted 13 January, 2021; v1 submitted 11 December, 2019;
originally announced December 2019.
-
High-Permittivity Pad Design Tool for 7T Neuroimaging and 3T Body Imaging
Authors:
J. H. F. van Gemert,
W. M. Brink,
A. G. Webb,
R. F. Remis
Abstract:
Purpose: High-permittivity materials in the form of flexible 'dielectric pads' have proved very useful for addressing RF inhomogeneities in high field MRI. Finding the optimal design of such pads is however a tedious task, reducing the impact of this technique. In this work we present an easy to use software tool which allows researchers and clinicians to design dielectric pads efficiently on stan…
▽ More
Purpose: High-permittivity materials in the form of flexible 'dielectric pads' have proved very useful for addressing RF inhomogeneities in high field MRI. Finding the optimal design of such pads is however a tedious task, reducing the impact of this technique. In this work we present an easy to use software tool which allows researchers and clinicians to design dielectric pads efficiently on standard computer systems, for 7T neuroimaging and 3T body imaging applications. Methods: The tool incorporates advanced computational methods based on field decomposition and model order reduction as a framework to efficiently evaluate the B1+ fields resulting from dielectric pads. The tool further incorporates an optimization routine to perform either straightforward design for one or two pads based on a target field approach, or a trade-off analysis between homogeneity and efficiency of the B1+ field in a specific region of interest. The 3T version further allows for shifting of the imaging landmark to enable different imaging targets to be centered in the body coil.Results: Example design results are shown for imaging the inner ear at 7T and for cardiac imaging at 3T. Computation times for all cases were under a couple of minutes. The change in homogeneity and efficiency with the pad's dimensions, location, and constitution are clearly listed. Conclusion: The developed tool can be easily used to design pads for any 7T neuroimaging and 3T body imaging application within minutes. This bridges the gap between the advanced design methods and the practical application by the MR community.
△ Less
Submitted 30 July, 2018;
originally announced July 2018.
