405905 research outputs found
Sort by
Efficient Epistemic Uncertainty Estimation in Cerebrovascular Segmentation
No full textBrain vessel segmentation of MR scans is a critical step in the diagnosis of cerebrovascular diseases. Due to the fine vessel structure, manual vessel segmentation is time consuming. Therefore, automatic deep learning (DL) based segmentation techniques are intensively investigated. As conventional DL models yield a high complexity and lack an indication of decision reliability, they are often considered as not trustworthy. This work aims to increase trust in DL based models by incorporating epistemic uncertainty quantification into cerebrovascular segmentation models for the first time. By implementing an efficient ensemble model combining the advantages of Bayesian Approximation and Deep Ensembles, we aim to overcome the high computational costs of conventional probabilistic networks. Areas of high model uncertainty and erroneous predictions are aligned which demonstrates the effectiveness and reliability of the approach. We perform extensive experiments applying the ensemble model on out-of-distribution (OOD) data. We demonstrate that for OOD-images, the estimated uncertainty increases. Additionally, omitting highly uncertain areas improves the segmentation quality, both for in- and out-of-distribution data. The ensemble model explains its limitations in a reliable manner and can maintain trustworthiness also for OOD data and could be considered in clinical application
Accelerating multigrid with streaming chiral SVD for Wilsonfermions in lattice QCD
No full textA modification to the setup algorithm for the multigrid preconditioner of Wilson fermions in lattice QCD is presented. A larger basis of test vectors than that used in conventional multigrid is calculated by the smoother and truncated by singular value decomposition on the chiral components of the test vectors. The truncated basis is used to form the prolongation and restriction matrices of the multigrid hierarchy. This modification of the setup method is demonstrated to increase the convergence of linear solvers on an anisotropic lattice with MeV from the Hadron Spectrum Collaboration and an isotropic lattice with MeV from the MILC Collaboration. The lattice volume dependence of the method is also examined. Increasing the number of test vectors improves speedup up to a point, but storing these vectors becomes impossible in limited memory resources such as GPUs. To address storage cost, we implement a \emph{streaming} singular value decomposition of the basis of test vectors on the chiral components and demonstrate a decrease in the number of fine level iterations by a factor of 1.7 for
Recent improvements in the detection of OH and RO₂ radicals and the development of compact and portable instrument for the measurement of OH reactivity
No full textGoing 3D with AI: Full 3D Reconstructions of Cytoarchitectonic Maps in BigBrain
No full textIntroduction:As part of the Julich-Brain Atlas (Amunts et al., 2020), the BigBrain dataset (Amunts et al., 2013) provides the first ultrahigh-resolution 3D model of the human brain at 20 μm isotropic resolution, reconstructed from 7,404 histological sections. It enables cytoarchitectonic mapping at a level of detail that bridges microscopic organization with macroscale imaging. Traditionally, areas have been delineated on 2D sections, limiting their integration into 3D brain reference spaces.Methods:We developed a hybrid workflow combining expert identification of cortical areas with deep learningbased 3D reconstruction. Using the ATLaSUI interface (Schiffer et al., 2021), neuroscientists annotated every 10th – 15th histological section, providing training data for the CytoNet model (manuscript in preparation). CytoNet infers cortical layer continuity and areal boundaries between annotated sections, avoiding geometric interpolation and preserving cytoarchitectonic detail. Largescale model inference and reconstruction were performed on the JURECA-DC supercomputer (Jülich Supercomputing Centre).Results:Currently, 33 cortical BigBrain areas are publicly available through the siibra tool suite and EBRAINS Knowledge Graph. An additional 56 cortical areas, including eight newly mapped regions, were reconstructed at 20 μm resolution, expanding the Julich-Brain Atlas to a total of 98 BigBrain areas. All maps will be openly accessible via EBRAINS, enabling interactive and programmatic exploration within a unified reference framework.Discussion:Since areas were reconstructed independently, minor overlaps can occur at region borders, especially in highly folded cortical zones. Sampling every 10th – 15th section may also introduce interpolation artifacts in regions with steep cytoarchitectonic transitions. Future work will focus on multi-area optimization to reduce boundary inconsistencies and improve 3D continuity. These ongoing developments advance the integration of microstructural and macroscale brain data within a coherent human brain reference space.Keywords: BigBrain, Julich Brain Atlas, Cytoarchitecture, 3D Reconstruction, Deep Learning, Human Brain Atlas, EBRAINSReferencesAmunts K, Lepage C, Borgeat L, et al. BigBrain: An ultrahigh-resolution 3D human brain model. Science. 2013;340(6139):1472-1475. doi:10.1126/science.1235381Amunts K, Mohlberg H, Bludau S, Zilles K. Julich-Brain: A 3D probabilistic atlas of the human brain’s cytoarchitecture. Science. 2020;369(6506):988-992. doi:10.1126/science.abb4588Schiffer C, Spitzer H, Kiwitz K, et al. Convolutional neural networks for cytoarchitectonic brain mapping at large scale. Neuroimage. 2021;240:118327. doi:10.1016/j.neuroimage.2021.118327Acknowledgments:This project received funding from the European Union’s Horizon 2020 Research and InnovationProgramme, grant agreement 101147319 (EBRAINS 2.0 Project), the Helmholtz Association portfolio theme “Supercomputing and Modeling for the Human Brain”, the Helmholtz Association’sInitiative and Networking Fund through the Helmholtz International BigBrain Analytics and LearningLaboratory (HIBALL) under the Helmholtz International Lab grant agreement InterLabs-0015, fromHELMHOLTZ IMAGING, a platform of the Helmholtz Information \& Data Science Incubator [XBRAIN, grant number: ZT-I-PF-4-061], and from the Deutsche Forschungsgemeinschaft (DFG,German Research Foundation) under the National Research Data Infrastructure – NFDI 46/1 –501864659
Closing the Gap: A Multi-Branch Diffusion Model for Inpainting at 1-µm Resolution
No full textAdvances in microscopic imaging and high-performance computing allow analyzing the complex cellular structure of the human brain in great detail. This progress has greatly aided in brain mapping and cell segmentation, and the development of automated analysis methods. However, histological image data can contain data gaps due to inevitable processing artifacts, which, despite careful precautions, may arise during histological lab work, such as missing sections, tissue tears, or inconsistent staining.To address this issue, we presented a convolutional neural network model that reconstructs corrupted data from surrounding tissue, while preserving precise cellular distributions. Our approach uses a denoising diffusion probabilistic model trained on light-microscopy scans of cell-body stained histological sections. We extended this model with the RePaint method to impute corrupted image data. We evaluate its performance with established deep learning models trained on the same type of histological data.A key challenge of our initial model was its difficulty in accurately reconstructing tissue boundaries and larger anatomical structures such as blood vessels. We address these challenges by an enhanced diffusion-based model that incorporates contextual information from adjacent sections of the brain. This model integrates three tissue patches from neighboring sections using a siamese network architecture with cross-attention mechanisms. Leveraging spatially aligned information across consecutive sections, our approach achieves a more anatomically coherent reconstruction.We demonstrate that our model significantly improves realism and anatomical plausibility of reconstructed cellular distributions, as measured by both cell density prediction and brain area classification tasks. The error in predicted cell density was reduced to below 5% across large inpainting regions, marking a notable improvement over previous approaches. In addition the model was evaluated on its ability to handle multiple missing sections at once, which resulted in no performance loss over the single missing section case. The model reliably preserves tissue borders and reconstructs larger structures like blood vessels, which are crucial for accurate cytoarchitectonic mapping.These findings underscore the potential of generative deep learning models for cytoarchitectonic research, opening new avenues for the automated reconstruction of histological data. Beyond inpainting small regions, our approach paves the way for the reconstruction of entirely missing brain slices, offering a powerful tool for bridging data gaps in high-resolution brain mapping efforts