Refine
Has Fulltext
- yes (4)
Is part of the Bibliography
- yes (4)
Year of publication
- 2023 (4) (remove)
Document Type
- Journal article (4) (remove)
Language
- English (4)
Keywords
- three-dimensional imaging (2)
- GRAPPA (1)
- RAKI (1)
- anatomy (1)
- biomedical engineering (1)
- bone imaging (1)
- complex‐valued machine learning (1)
- data augmentation (1)
- deep learning (1)
- diagnosis (1)
- electrical and electronic engineering (1)
- halogenation (1)
- imaging (1)
- medical imaging (1)
- non‐fullerene acceptors (1)
- organic photovoltaics (1)
- parallel imaging (1)
- software (1)
- spin physics (1)
- tomography (1)
- triplet excitons (1)
Institute
- Physikalisches Institut (4) (remove)
Automated analysis of the inner ear anatomy in radiological data instead of time-consuming manual assessment is a worthwhile goal that could facilitate preoperative planning and clinical research. We propose a framework encompassing joint semantic segmentation of the inner ear and anatomical landmark detection of helicotrema, oval and round window. A fully automated pipeline with a single, dual-headed volumetric 3D U-Net was implemented, trained and evaluated using manually labeled in-house datasets from cadaveric specimen (N = 43) and clinical practice (N = 9). The model robustness was further evaluated on three independent open-source datasets (N = 23 + 7 + 17 scans) consisting of cadaveric specimen scans. For the in-house datasets, Dice scores of 0.97 and 0.94, intersection-over-union scores of 0.94 and 0.89 and average Hausdorf distances of 0.065 and 0.14 voxel units were achieved. The landmark localization task was performed automatically with an average localization error of 3.3 and 5.2 voxel units. A robust, albeit reduced performance could be
attained for the catalogue of three open-source datasets. Results of the ablation studies with 43 mono-parametric variations of the basal architecture and training protocol provided task-optimal parameters for both categories. Ablation studies against single-task variants of the basal architecture showed a clear performance beneft of coupling landmark localization with segmentation and a dataset-dependent performance impact on segmentation ability.
Minimally invasive endovascular interventions have become an important tool for the treatment of cardiovascular diseases such as ischemic heart disease, peripheral artery disease, and stroke. X-ray fluoroscopy and digital subtraction angiography are used to precisely guide these procedures, but they are associated with radiation exposure for patients and clinical staff. Magnetic Particle Imaging (MPI) is an emerging imaging technology using time-varying magnetic fields combined with magnetic nanoparticle tracers for fast and highly sensitive imaging. In recent years, basic experiments have shown that MPI has great potential for cardiovascular applications. However, commercially available MPI scanners were too large and expensive and had a small field of view (FOV) designed for rodents, which limited further translational research. The first human-sized MPI scanner designed specifically for brain imaging showed promising results but had limitations in gradient strength, acquisition time and portability. Here, we present a portable interventional MPI (iMPI) system dedicated for real-time endovascular interventions free of ionizing radiation. It uses a novel field generator approach with a very large FOV and an application-oriented open design enabling hybrid approaches with conventional X-ray-based angiography. The feasibility of a real-time iMPI-guided percutaneous transluminal angioplasty (PTA) is shown in a realistic dynamic human-sized leg model.
To evaluate an iterative learning approach for enhanced performance of robust artificial‐neural‐networks for k‐space interpolation (RAKI), when only a limited amount of training data (auto‐calibration signals [ACS]) are available for accelerated standard 2D imaging.
Methods
In a first step, the RAKI model was tailored for the case of limited training data amount. In the iterative learning approach (termed iterative RAKI [iRAKI]), the tailored RAKI model is initially trained using original and augmented ACS obtained from a linear parallel imaging reconstruction. Subsequently, the RAKI convolution filters are refined iteratively using original and augmented ACS extracted from the previous RAKI reconstruction. Evaluation was carried out on 200 retrospectively undersampled in vivo datasets from the fastMRI neuro database with different contrast settings.
Results
For limited training data (18 and 22 ACS lines for R = 4 and R = 5, respectively), iRAKI outperforms standard RAKI by reducing residual artifacts and yields better noise suppression when compared to standard parallel imaging, underlined by quantitative reconstruction quality metrics. Additionally, iRAKI shows better performance than both GRAPPA and standard RAKI in case of pre‐scan calibration with varying contrast between training‐ and undersampled data.
Conclusion
RAKI benefits from the iterative learning approach, which preserves the noise suppression feature, but requires less original training data for the accurate reconstruction of standard 2D images thereby improving net acceleration.
The great progress in organic photovoltaics (OPV) over the past few years has been largely achieved by the development of non‐fullerene acceptors (NFAs), with power conversion efficiencies now approaching 20%. To further improve device performance, loss mechanisms must be identified and minimized. Triplet states are known to adversely affect device performance, since they can form energetically trapped excitons on low‐lying states that are responsible for non‐radiative losses or even device degradation. Halogenation of OPV materials has long been employed to tailor energy levels and to enhance open circuit voltage. Yet, the influence on recombination to triplet excitons has been largely unexplored. Using the complementary spin‐sensitive methods of photoluminescence detected magnetic resonance and transient electron paramagnetic resonance corroborated by transient absorption and quantum‐chemical calculations, exciton pathways in OPV blends are unravelled employing the polymer donors PBDB‐T, PM6, and PM7 together with NFAs Y6 and Y7. All blends reveal triplet excitons on the NFA populated via non‐geminate hole back transfer and, in blends with halogenated donors, also by spin‐orbit coupling driven intersystem crossing. Identifying these triplet formation pathways in all tested solar cell absorber films highlights the untapped potential for improved charge generation to further increase plateauing OPV efficiencies.