Search

We present Connected-Component~(CC)-Metrics, a novel semantic segmentation evaluation protocol, targeted to align existing semantic segmentation metrics to a multi-instance detection scenario in which each connected component matters. We motivate this setup in the common medical scenario of semantic metastases segmentation in a full-body PET/CT. We show how existing semantic segmentation metrics suffer from a bias towards larger connected components contradicting the clinical assessment of scans in which tumor size and clinical relevance are uncorrelated. To rebalance existing segmentation metrics, we propose to evaluate them on a per-component basis thus giving each tumor the same weight irrespective of its size. To match predictions to ground-truth segments, we employ a proximity-based matching criterion, evaluating common metrics locally at the component of interest. Using this approach, we break free of biases introduced by large metastasis for overlap-based metrics such as Dice or Surface Dice. CC-Metrics also improves distance-based metrics such as Hausdorff Distances which are uninformative for small changes that do not influence the maximum or 95th percentile, and avoids pitfalls introduced by directly combining counting-based metrics with overlap-based metrics as it is done in Panoptic Quality.

Medical data employed in research frequently comprises sensitive patient health information (PHI), which is subject to rigorous legal frameworks such as the General Data Protection Regulation (GDPR) or the Health Insurance Portability and Accountability Act (HIPAA). Consequently, these types of data must be pseudonymized prior to utilisation, which presents a significant challenge for many researchers. Given the vast array of medical data, it is necessary to employ a variety of de-identification techniques. To facilitate the anonymization process for medical imaging data, we have developed an open-source tool that can be used to de-identify DICOM magnetic resonance images, computer tomography images, whole slide images and magnetic resonance twix raw data. Furthermore, the implementation of a neural network enables the removal of text within the images. The proposed tool automates an elaborate anonymization pipeline for multiple types of inputs, reducing the need for additional tools used for de-identification of imaging data. We make our code publicly available at https://github.com/code-lukas/medical_image_deidentification.

Unstructured data in industries such as healthcare, finance, and manufacturing presents significant challenges for efficient analysis and decision making. Detecting patterns within this data and understanding their impact is critical but complex without the right tools. Traditionally, these tasks relied on the expertise of data analysts or labor-intensive manual reviews. In response, we introduce Spacewalker, an interactive tool designed to explore and annotate data across multiple modalities. Spacewalker allows users to extract data representations and visualize them in low-dimensional spaces, enabling the detection of semantic similarities. Through extensive user studies, we assess Spacewalker's effectiveness in data annotation and integrity verification. Results show that the tool's ability to traverse latent spaces and perform multi-modal queries significantly enhances the user's capacity to quickly identify relevant data. Moreover, Spacewalker allows for annotation speed-ups far superior to conventional methods, making it a promising tool for efficiently navigating unstructured data and improving decision making processes. The code of this work is open-source and can be found at: https://github.com/code-lukas/Spacewalker

Recognizing pain in video is crucial for improving patient-computer interaction systems, yet traditional data collection in this domain raises significant ethical and logistical challenges. This study introduces a novel approach that leverages synthetic data to enhance video-based pain recognition models, providing an ethical and scalable alternative. We present a pipeline that synthesizes realistic 3D facial models by capturing nuanced facial movements from a small participant pool, and mapping these onto diverse synthetic avatars. This process generates 8,600 synthetic faces, accurately reflecting genuine pain expressions from varied angles and perspectives. Utilizing advanced facial capture techniques, and leveraging public datasets like CelebV-HQ and FFHQ-UV for demographic diversity, our new synthetic dataset significantly enhances model training while ensuring privacy by anonymizing identities through facial replacements. Experimental results demonstrate that models trained on combinations of synthetic data paired with a small amount of real participants achieve superior performance in pain recognition, effectively bridging the gap between synthetic simulations and real-world applications. Our approach addresses data scarcity and ethical concerns, offering a new solution for pain detection and opening new avenues for research in privacy-preserving dataset generation. All resources are publicly available to encourage further innovation in this field., Comment: Pain Recognition Synthetic Data Video Analysis Privacy Preserving

Lesion segmentation in PET/CT imaging is essential for precise tumor characterization, which supports personalized treatment planning and enhances diagnostic precision in oncology. However, accurate manual segmentation of lesions is time-consuming and prone to inter-observer variability. Given the rising demand and clinical use of PET/CT, automated segmentation methods, particularly deep-learning-based approaches, have become increasingly more relevant. The autoPET III Challenge focuses on advancing automated segmentation of tumor lesions in PET/CT images in a multitracer multicenter setting, addressing the clinical need for quantitative, robust, and generalizable solutions. Building on previous challenges, the third iteration of the autoPET challenge introduces a more diverse dataset featuring two different tracers (FDG and PSMA) from two clinical centers. To this extent, we developed a classifier that identifies the tracer of the given PET/CT based on the Maximum Intensity Projection of the PET scan. We trained two individual nnUNet-ensembles for each tracer where anatomical labels are included as a multi-label task to enhance the model's performance. Our final submission achieves cross-validation Dice scores of 76.90% and 61.33% for the publicly available FDG and PSMA datasets, respectively. The code is available at https://github.com/hakal104/autoPETIII/ ., Comment: AutoPET III challenge submission

Motivated by the recent advances in the understanding of integrability for $AdS_3$ backgrounds, we present a lightning review of this approach, with particular attention to the "tensionless" limits (with zero and one unit of NSNS flux), and to the many integrable deformations of the supergravity backgrounds. Our aim is to concisely but comprehensively take stock of the state of the art in the field, in a way accessible to non-experts, and to highlight outstanding challenges. Along the way we reference where the various derivation of these results, which we mostly omit, can be found in full detail., Comment: Review prepared for the MATRIX Annals; 52 pages, 4 figures (3 of which from 2308.11576); feedback welcome

Pathological structures in medical images are typically deviations from the expected anatomy of a patient. While clinicians consider this interplay between anatomy and pathology, recent deep learning algorithms specialize in recognizing either one of the two, rarely considering the patient's body from such a joint perspective. In this paper, we develop a generalist segmentation model that combines anatomical and pathological information, aiming to enhance the segmentation accuracy of pathological features. Our Anatomy-Pathology Exchange (APEx) training utilizes a query-based segmentation transformer which decodes a joint feature space into query-representations for human anatomy and interleaves them via a mixing strategy into the pathology-decoder for anatomy-informed pathology predictions. In doing so, we are able to report the best results across the board on FDG-PET-CT and Chest X-Ray pathology segmentation tasks with a margin of up to 3.3% as compared to strong baseline methods. Code and models will be publicly available at github.com/alexanderjaus/APEx.

We present a novel preprocessing and prediction pipeline for the classification of magnetic resonance imaging (MRI) that takes advantage of the information rich complex valued k-Space. Using a publicly available MRI raw dataset with 312 subject and a total of 9508 slices, we show the advantage of utilizing the k-Space for better prostate cancer likelihood estimation in comparison to just using the magnitudinal information in the image domain, with an AUROC of $86.1\%\pm1.8\%$. Additionally, by using high undersampling rates and a simple principal component analysis (PCA) for coil compression, we reduce the time needed for reconstruction by avoiding the time intensive GRAPPA reconstruction algorithm. By using digital undersampling for our experiments, we show that scanning and reconstruction time could be reduced. Even with an undersampling factor of 16, our approach achieves meaningful results, with an AUROC of $71.4\%\pm2.9\%$, using the PCA coil combination and taking into account the k-Space information. With this study, we were able to show the feasibility of preserving phase and k-Space information, with consistent results. Besides preserving valuable information for further diagnostics, this approach can work without the time intensive ADC and reconstruction calculations, greatly reducing the post processing, as well as potential scanning time, increasing patient comfort and allowing a close to real-time prediction.

Previous controlled experiments on single-lane ring roads have shown that a single partially autonomous vehicle (AV) can effectively mitigate traffic waves. This naturally prompts the question of how these findings can be generalized to field operational, high-density traffic conditions. To address this question, the Congestion Impacts Reduction via CAV-in-the-loop Lagrangian Energy Smoothing (CIRCLES) Consortium conducted MegaVanderTest (MVT), a live traffic control experiment involving 100 vehicles near Nashville, TN, USA. This article is a tutorial for developing analytical and simulation-based tools essential for designing and executing a live traffic control experiment like the MVT. It presents an overview of the proposed roadmap and various procedures used in designing, monitoring, and conducting the MVT, which is the largest mobile traffic control experiment at the time. The design process is aimed at evaluating the impact of the CIRCLES AVs on surrounding traffic. The article discusses the agent-based traffic simulation framework created for this evaluation. A novel methodological framework is introduced to calibrate this microsimulation, aiming to accurately capture traffic dynamics and assess the impact of adding 100 vehicles to existing traffic. The calibration model's effectiveness is verified using data from a six-mile section of Nashville's I-24 highway. The results indicate that the proposed model establishes an effective feedback loop between the optimizer and the simulator, thereby calibrating flow and speed with different spatiotemporal characteristics to minimize the error between simulated and real-world data. Finally, We simulate AVs in multiple scenarios to assess their effect on traffic congestion. This evaluation validates the AV routes, thereby contributing to the execution of a safe and successful live traffic control experiment via AVs.

Periodically-driven systems engender a rich competition between the time scales of the drives and those of the system, leading to a limited ability to describe the system in full. We present a framework for the description of interacting bosonic driven systems via a Floquet expansion on top of a quantization that "counts" the drive photons, and provide compelling arguments for the superior performance of our method relative to standard Floquet approaches. Crucially, our approach extends beyond the rotating wave approximation and addresses the long-standing issue of mismatch between the quantum Floquet formalism and its classical counterpart. We, furthermore, pinpoint key corrections to the positions of multiphoton resonances, which are commonly used in the calibration and operation of qubit architectures., Comment: 5 pages, 3 figures, comments are welcome

We compute the one-loop $2 \rightarrow 2$ scattering amplitude of massless scalars on the world volume of an infinite $D = 11$ supermembrane quantized in the static gauge. The resulting expression is manifestly finite and turns out to be much simpler than in the bosonic membrane case in arXiv:2308.12189 being simply proportional to the tree-level scattering amplitude. We also consider the case of $\mathbb{R}^{1,1} \times S^1$ membrane with one dimension compactified on a circle of radius R and demonstrate how the supermembrane scattering amplitude reduces to the one on an infinite $D = 10$ Green-Schwarz superstring in the limit of R $\rightarrow 0$., Comment: 28 pages, v3

Large Language Models (LLMs) are expected to significantly contribute to patient care, diagnostics, and administrative processes. Emerging biomedical LLMs aim to address healthcare-specific challenges, including privacy demands and computational constraints. Assessing the models' suitability for this sensitive application area is of the utmost importance. However, biomedical training has not been systematically evaluated on medical tasks. This study investigates the effect of biomedical training in the context of six practical medical tasks evaluating $25$ models. In contrast to previous evaluations, our results reveal a performance decline in nine out of twelve biomedical models after fine-tuning, particularly on tasks involving hallucinations, ICD10 coding, and instruction adherence. General-domain models like Meta-Llama-3.1-70B-Instruct outperformed their biomedical counterparts, indicating a trade-off between domain-specific fine-tuning and general medical task performance. We open-source all evaluation scripts and datasets at https://github.com/TIO-IKIM/CLUE to support further research in this critical area.

In this article, we explore the technical details of the reinforcement learning (RL) algorithms that were deployed in the largest field test of automated vehicles designed to smooth traffic flow in history as of 2023, uncovering the challenges and breakthroughs that come with developing RL controllers for automated vehicles. We delve into the fundamental concepts behind RL algorithms and their application in the context of self-driving cars, discussing the developmental process from simulation to deployment in detail, from designing simulators to reward function shaping. We present the results in both simulation and deployment, discussing the flow-smoothing benefits of the RL controller. From understanding the basics of Markov decision processes to exploring advanced techniques such as deep RL, our article offers a comprehensive overview and deep dive of the theoretical foundations and practical implementations driving this rapidly evolving field. We also showcase real-world case studies and alternative research projects that highlight the impact of RL controllers in revolutionizing autonomous driving. From tackling complex urban environments to dealing with unpredictable traffic scenarios, these intelligent controllers are pushing the boundaries of what automated vehicles can achieve. Furthermore, we examine the safety considerations and hardware-focused technical details surrounding deployment of RL controllers into automated vehicles. As these algorithms learn and evolve through interactions with the environment, ensuring their behavior aligns with safety standards becomes crucial. We explore the methodologies and frameworks being developed to address these challenges, emphasizing the importance of building reliable control systems for automated vehicles.

The CIRCLES project aims to reduce instabilities in traffic flow, which are naturally occurring phenomena due to human driving behavior. These "phantom jams" or "stop-and-go waves,"are a significant source of wasted energy. Toward this goal, the CIRCLES project designed a control system referred to as the MegaController by the CIRCLES team, that could be deployed in real traffic. Our field experiment leveraged a heterogeneous fleet of 100 longitudinally-controlled vehicles as Lagrangian traffic actuators, each of which ran a controller with the architecture described in this paper. The MegaController is a hierarchical control architecture, which consists of two main layers. The upper layer is called Speed Planner, and is a centralized optimal control algorithm. It assigns speed targets to the vehicles, conveyed through the LTE cellular network. The lower layer is a control layer, running on each vehicle. It performs local actuation by overriding the stock adaptive cruise controller, using the stock on-board sensors. The Speed Planner ingests live data feeds provided by third parties, as well as data from our own control vehicles, and uses both to perform the speed assignment. The architecture of the speed planner allows for modular use of standard control techniques, such as optimal control, model predictive control, kernel methods and others, including Deep RL, model predictive control and explicit controllers. Depending on the vehicle architecture, all onboard sensing data can be accessed by the local controllers, or only some. Control inputs vary across different automakers, with inputs ranging from torque or acceleration requests for some cars, and electronic selection of ACC set points in others. The proposed architecture allows for the combination of all possible settings proposed above. Most configurations were tested throughout the ramp up to the MegaVandertest.

This paper introduces a novel control framework for Lagrangian variable speed limits in hybrid traffic flow environments utilizing automated vehicles (AVs). The framework was validated using a fleet of 100 connected automated vehicles as part of the largest coordinated open-road test designed to smooth traffic flow. The framework includes two main components: a high-level controller deployed on the server side, named Speed Planner, and low-level controllers called vehicle controllers deployed on the vehicle side. The Speed Planner designs and updates target speeds for the vehicle controllers based on real-time Traffic State Estimation (TSE) [1]. The Speed Planner comprises two modules: a TSE enhancement module and a target speed design module. The TSE enhancement module is designed to minimize the effects of inherent latency in the received traffic information and to improve the spatial and temporal resolution of the input traffic data. The target speed design module generates target speed profiles with the goal of improving traffic flow. The vehicle controllers are designed to track the target speed meanwhile responding to the surrounding situation. The numerical simulation indicates the performance of the proposed method: the bottleneck throughput has increased by 5.01%, and the speed standard deviation has been reduced by a significant 34.36%. We further showcase an operational study with a description of how the controller was implemented on a field-test with 100 AVs and its comprehensive effects on the traffic flow.

Light-cone gauge-fixed sigma-models on $AdS_n\times S^n$ backgrounds play an important role in the integrability formulation of the AdS/CFT correspondence. The string spectrum of the sigma-model is gauge-independent, however the Hamiltonian and scattering matrix of the transverse worldsheet fields are not. We study how these change for a large family of inequivalent light-cone gauges, which are interpreted as $T\bar{T}$, $\tilde{J}T_\tau$, $JT_\sigma$ and $J^\tau$ deformations. We investigate the moduli space of equivalent light-cone gauges and, specialising to $AdS_5 \times S^5$, compute the different light-cone gauge symmetry algebras, well-known to be $\mathfrak{psu}(2|2)^{\oplus 2} \oplus \mathfrak{u}(1)^{\oplus 2}$ for the standard gauge-fixing. Many integrable deformations require a non-standard light-cone gauge, hence our classification and analysis of inequivalent gauges will be important for analysing such models., Comment: 52 pages, 3 tables; v2: expanded discussion on moduli space and pp-wave analysis (sec. 3)

With the aim of investigating the existence of an integrable elliptic deformation of strings on $\mathsf{AdS}_3 \times \mathsf{S}^3 \times \mathsf{T}^4$, we compute the tree-level worldsheet S-matrix of the elliptically-deformed bosonic sigma model on $\mathsf{AdS}_3 \times \mathsf{S}^3$ in uniform light-cone gauge. The resulting tree-level S-matrix is compatible with the integrability of the model and has interesting features, including a hidden $\mathsf{U}(1)$ symmetry not manifest in the Lagrangian. We find that it cannot be embedded in the known exact integrable R-matrices describing deformations of the undeformed $\mathsf{AdS}_3 \times \mathsf{S}^3 \times \mathsf{T}^4$ light-cone gauge S-matrix including fermions. Therefore, we construct embeddings of the deformed 6-d metric in type II supergravity with constant dilaton and homogeneous fluxes. The simplicity of these solutions suggests they are promising candidates to lead to an integrable string sigma model including fermions., Comment: 28 pages; v2 minor comments added, matches published version

We review different exact approaches to string theory. In the context of the Green-Schwarz superstring, we discuss the action in curved backgrounds and its supercoset formulation, with particular attention to superstring backgrounds of the $AdS_3$ type supported by both Ramond-Ramond and Neveu-Schwarz-Neveu-Schwarz fluxes. This is the basis for the discussion of classical integrability, of worldsheet-scattering factorisation in the uniform lightcone gauge, and eventually of the string spectrum through the mirror thermodynamic Bethe ansatz, which for $AdS_3$ backgrounds was only derived and analysed very recently. We then illustrate some aspects of the Ramond-Neveu-Schwarz string, and introduce the formalism of Berkovits-Vafa-Witten, which has seen very recent applications to $AdS_3$ physics, which we also briefly review. Finally, we present the relation between M-theory in the discrete lightcone quantisation and decoupling limits of string theory that exhibit non-relativistic behaviours, highlighting the connection with integrable $T\bar{T}$ deformations, as well as the relation between spin-matrix theory and Landau-Lifshitz models. This review is based on lectures given at the Young Researchers Integrability School and Workshop 2022 "Taming the string worldsheet" at NORDITA, Stockholm., Comment: 283 pages; v2: references added

Automatic generation of morphed face images often produces ghosting artifacts due to poorly aligned structures in the input images. Manual processing can mitigate these artifacts. However, this is not feasible for the generation of large datasets, which are required for training and evaluating robust morphing attack detectors. In this paper, we propose a method for automatic prevention of ghosting artifacts based on a pixel-wise alignment during morph generation. We evaluate our proposed method on state-of-the-art detectors and show that our morphs are harder to detect, particularly, when combined with style-transfer-based improvement of low-level image characteristics. Furthermore, we show that our approach does not impair the biometric quality, which is essential for high quality morphs., Comment: Accepted at VISAPP 2024

Recently developed terahertz (THz) two-dimensional coherent spectroscopy (2DCS) is a powerful technique to obtain materials information in a fashion qualitatively different from other spectroscopies. Here, we utilized THz 2DCS to investigate the THz nonlinear response of conventional superconductor NbN. Using broad-band THz pulses as light sources, we observed a third-order nonlinear signal whose spectral components are peaked at twice the superconducting gap energy $2\Delta$. With narrow-band THz pulses, a THz nonlinear signal was identified at the driving frequency $\Omega$ and exhibited a resonant enhancement at temperature when $\Omega = 2\Delta$. General theoretical considerations show that such a resonance can only arise from a disorder-activated paramagnetic coupling between the light and the electronic current. This proves that the nonlinear THz response can access processes distinct from the diamagnetic Raman-like density fluctuations, which are believed to dominate the nonlinear response at optical frequencies in metals. Our numerical simulations reveal that even for a small amount of disorder, the $\Omega=2\Delta$ resonance is dominated by the superconducting amplitude mode over the entire investigated disorder range. This is in contrast to other resonances, whose amplitude-mode contribution depends on disorder. Our findings demonstrate the unique ability of THz 2DCS to explore collective excitations inaccessible in other spectroscopies.

The term Research Software Engineer, or RSE, emerged a little over 10 years ago as a way to represent individuals working in the research community but focusing on software development. The term has been widely adopted and there are a number of high-level definitions of what an RSE is. However, the roles of RSEs vary depending on the institutional context they work in. At one end of the spectrum, RSE roles may look similar to a traditional research role. At the other extreme, they resemble that of a software engineer in industry. Most RSE roles inhabit the space between these two extremes. Therefore, providing a straightforward, comprehensive definition of what an RSE does and what experience, skills and competencies are required to become one is challenging. In this community paper we define the broad notion of what an RSE is, explore the different types of work they undertake, and define a list of fundamental competencies as well as values that define the general profile of an RSE. On this basis, we elaborate on the progression of these skills along different dimensions, looking at specific types of RSE roles, proposing recommendations for organisations, and giving examples of future specialisations. An appendix details how existing curricula fit into this framework., Comment: 34 pages, public repository for feedback here: https://github.com/the-teachingRSE-project/competencies

We demonstrate a new capability of automated vehicles: mixed autonomy traffic control. With this new capability, automated vehicles can shape the traffic flows composed of other non-automated vehicles, which has the promise to improve safety, efficiency, and energy outcomes in transportation systems at a societal scale. Investigating mixed autonomy mobile traffic control must be done in situ given that the complex dynamics of other drivers and their response to a team of automated vehicles cannot be effectively modeled. This capability has been blocked because there is no existing scalable and affordable platform for experimental control. This paper introduces an extensible open-source hardware and software platform, enabling a team of 100 vehicles to execute several different vehicular control algorithms as a collaborative fleet, composed of three different makes and models, which drove 22752 miles in a combined 1022 hours, over 5 days in Nashville, TN in November 2022.

The dissipation of stop-and-go waves attracted recent attention as a traffic management problem, which can be efficiently addressed by automated driving. As part of the 100 automated vehicles experiment named MegaVanderTest, feedback controls were used to induce strong dissipation via velocity smoothing. More precisely, a single vehicle driving differently in one of the four lanes of I-24 in the Nashville area was able to regularize the velocity profile by reducing oscillations in time and velocity differences among vehicles. Quantitative measures of this effect were possible due to the innovative I-24 MOTION system capable of monitoring the traffic conditions for all vehicles on the roadway. This paper presents the control design, the technological aspects involved in its deployment, and, finally, the results achieved by the experiment., Comment: 21 pages, 1 Table , 9 figures

Layered cuprates offer a preferential playground for optical non-linearity thanks to the emergence, below Tc, of soft out-of-plane Josephson plasmons. The hallmark of such a non-linearity is the observation of Third Harmonic Generation, that has been theoretically understood as a sum-frequency process involving a two-plasmon excitation. However, recent experiments in cuprates with two planes per unit cell challenge this interpretation, due to the lack of resonant response at the temperature where the driving frequency matches the plasma energy scale, as observed instead in single-layer cuprates. Here we show that such an apparent discrepancy in bilayer systems can be resolved by taking into account the combined effect of light polarization and Josephson-coupling anisotropy on setting the energy range where three-dimensional layered plasma modes can be resonantly excited. Our results offer a novel perspective on the possibility to tune on demand high-harmonic generation by artificially designing Josephson heterostructures.

Traditionally, large language models have been either trained on general web crawls or domain-specific data. However, recent successes of generative large language models, have shed light on the benefits of cross-domain datasets. To examine the significance of prioritizing data diversity over quality, we present a German dataset comprising texts from five domains, along with another dataset aimed at containing high-quality data. Through training a series of models ranging between 122M and 750M parameters on both datasets, we conduct a comprehensive benchmark on multiple downstream tasks. Our findings demonstrate that the models trained on the cross-domain dataset outperform those trained on quality data alone, leading to improvements up to $4.45\%$ over the previous state-of-the-art. The models are available at https://huggingface.co/ikim-uk-essen, Comment: 13 pages, 1 figure, accepted at Findings of the Association for Computational Linguistics: EMNLP 2023

The energy demand of vehicles, particularly in unsteady drive cycles, is affected by complex dynamics internal to the engine and other powertrain components. Yet, in many applications, particularly macroscopic traffic flow modeling and optimization, structurally simple approximations to the complex vehicle dynamics are needed that nevertheless reproduce the correct effective energy behavior. This work presents a systematic model reduction pipeline that starts from complex vehicle models based on the Autonomie software and derives a hierarchy of simplified models that are fast to evaluate, easy to disseminate in open-source frameworks, and compatible with optimization frameworks. The pipeline, based on a virtual chassis dynamometer and subsequent approximation strategies, is reproducible and is applied to six different vehicle classes to produce concrete explicit energy models that represent an average vehicle in each class and leverage the accuracy and validation work of the Autonomie software., Comment: 40 pages, 9 figures

Traffic waves can rise even from single lane car-following behaviour. To better understand and mitigate traffic waves, it is necessary to use analytical tools like mathematical models, data analysis, and micro-simulations that can capture the dynamics of real traffic flow. In this study, we isolate car-following dynamics and present a systematic hierarchy of tests that connect the microscopic scale with the meaningful macroscopic effective state in the presence of waves. This allows insights with precise attributable cause-to-effect relationships of specific observed traffic patterns. We establish a principled way of generating macroscopic flow quantities from microscopic models in the unstable regime. Those quantities are then used to study how the corresponding non-equilibrium wave structures manifest in the fundamental diagram, based on three basic scenarios that can serve as building blocks for understanding more complex micro-simulation studies. Finally, this study gives insight on the shapes of the reduced fundamental diagrams for different commonly used microscopic models., Comment: 22 pages, 12 figures

Explicit Runge--Kutta (RK) methods are susceptible to a reduction in the observed order of convergence when applied to initial-boundary value problem with time-dependent boundary conditions. We study conditions on explicit RK methods that guarantee high-order convergence for linear problems; we refer to these conditions as weak stage order conditions. We prove a general relationship between the method's order, weak stage order, and number of stages. We derive explicit RK methods with high weak stage order and demonstrate, through numerical tests, that they avoid the order reduction phenomenon up to any order for linear problems and up to order three for nonlinear problems.

This work introduces a multi-camera tracking dataset consisting of 234 hours of video data recorded concurrently from 234 overlapping HD cameras covering a 4.2 mile stretch of 8-10 lane interstate highway near Nashville, TN. The video is recorded during a period of high traffic density with 500+ objects typically visible within the scene and typical object longevities of 3-15 minutes. GPS trajectories from 270 vehicle passes through the scene are manually corrected in the video data to provide a set of ground-truth trajectories for recall-oriented tracking metrics, and object detections are provided for each camera in the scene (159 million total before cross-camera fusion). Initial benchmarking of tracking-by-detection algorithms is performed against the GPS trajectories, and a best HOTA of only 9.5% is obtained (best recall 75.9% at IOU 0.1, 47.9 average IDs per ground truth object), indicating the benchmarked trackers do not perform sufficiently well at the long temporal and spatial durations required for traffic scene understanding.

Prior to the deep learning era, shape was commonly used to describe the objects. Nowadays, state-of-the-art (SOTA) algorithms in medical imaging are predominantly diverging from computer vision, where voxel grids, meshes, point clouds, and implicit surface models are used. This is seen from numerous shape-related publications in premier vision conferences as well as the growing popularity of ShapeNet (about 51,300 models) and Princeton ModelNet (127,915 models). For the medical domain, we present a large collection of anatomical shapes (e.g., bones, organs, vessels) and 3D models of surgical instrument, called MedShapeNet, created to facilitate the translation of data-driven vision algorithms to medical applications and to adapt SOTA vision algorithms to medical problems. As a unique feature, we directly model the majority of shapes on the imaging data of real patients. As of today, MedShapeNet includes 23 dataset with more than 100,000 shapes that are paired with annotations (ground truth). Our data is freely accessible via a web interface and a Python application programming interface (API) and can be used for discriminative, reconstructive, and variational benchmarks as well as various applications in virtual, augmented, or mixed reality, and 3D printing. Exemplary, we present use cases in the fields of classification of brain tumors, facial and skull reconstructions, multi-class anatomy completion, education, and 3D printing. In future, we will extend the data and improve the interfaces. The project pages are: https://medshapenet.ikim.nrw/ and https://github.com/Jianningli/medshapenet-feedback, Comment: 16 pages

Expanding Nambu-Goto action near infinitely long string vacuum one can compute scattering amplitudes of 2d massless fields representing transverse string coordinates. As was shown in arXiv:1203.1054, the resulting S-matrix is integrable, in agreement with the known free string spectrum and also with an interpretation of the static-gauge NG action as a $T\bar T$ deformation of a free massless theory. We consider a generalization of this computation to the case of a membrane, expanding its 3d action near an infinite membrane vacuum that has cylindrical $\mathbb R \times S^1$ shape (we refer to such membrane as "compactified"). Representing 3d fields as Fourier series in $S^1$ coordinate we get an effective 2d model in which the massless string modes are coupled to an infinite KK tower of massive 2d modes. We find that the resulting 2d S-matrix is not integrable already at the tree level. We also compute 1-loop scattering amplitude of massless string modes with all compactified membrane modes propagating in the loop. The result is UV finite and is a non-trivial function of the kinematic variables. In the large momentum limit or when the radius of $S^1$ is taken to infinity we recover the expression for the 1-loop scattering amplitude of the uncompactified $\mathbb R^2$ membrane. We also consider a 2d model which is the $T\bar T$ deformation to the free theory with the same massless plus infinite massive tower of modes. The corresponding 2d S-matrix is found, as expected, to be integrable., Comment: 34 pages, v3

This paper develops low-dimensional moment methods for advective problems on networks of domains. The evolution of a density function is described by a linear advection-diffusion-reaction equation on each domain, combined via advective flux coupling across domains in the network graph. The PDEs' coefficients vary in time and across domains but they are fixed along each domain. As a result, the solution on each domain is frequently close to a Gaussian that moves, decays, and widens. For that reason, this work studies moment methods that track only three degrees of freedom per domain -- in contrast to traditional PDE discretization methods that tend to require many more variables per domain. A simple ODE-based moment method is developed, as well as an asymptotic-preserving scheme. We apply the methodology to an application that models the life cycle of forest pests that undergo different life stages and developmental pathways. The model is calibrated for the spotted lanternfly, an invasive species present in the Eastern USA. We showcase that the moment method, despite its significant low-dimensionality, can successfully reproduce the prediction of the pest's establishment potential, compared to much higher-dimensional computational approaches., Comment: 31 pages, 14 figures

In this study, we present a method for generating automated anatomy segmentation datasets using a sequential process that involves nnU-Net-based pseudo-labeling and anatomy-guided pseudo-label refinement. By combining various fragmented knowledge bases, we generate a dataset of whole-body CT scans with $142$ voxel-level labels for 533 volumes providing comprehensive anatomical coverage which experts have approved. Our proposed procedure does not rely on manual annotation during the label aggregation stage. We examine its plausibility and usefulness using three complementary checks: Human expert evaluation which approved the dataset, a Deep Learning usefulness benchmark on the BTCV dataset in which we achieve 85% dice score without using its training dataset, and medical validity checks. This evaluation procedure combines scalable automated checks with labor-intensive high-quality expert checks. Besides the dataset, we release our trained unified anatomical segmentation model capable of predicting $142$ anatomical structures on CT data., Comment: 18 pages, 8 figures, 2 tables

We present a detailed theoretical investigation of the linear and non-linear optical response in a model system for a disordered d-wave superconductor, showing that for both quantities the gap symmetry considerably changes the paradigm of the optical response based on the conventional s-wave case. For what concerns the linear response our findings agree with previous work showing that in strongly-disordered d-wave superconductors a large fraction of uncondensed spectral weight survives below Tc, making the optical absorption around the gap-frequency scale almost unchanged with respect to the normal state. Our numerical results are in excellent quantitative agreement with experiments in overdoped cuprates. In the non-linear regime we focus on the third-harmonic generation (THG), finding that, as already established for the s-wave case, in general a large THG is triggered by disorder-activated paramagnetic processes. However, in the d-wave case the BCS response is monotonously increasing in frequency, loosing any signature of THG enhancement when the THz pump frequency $\omega$ matches the gap maximum $\Delta$, a hallmark of previous experiments in conventional s-wave superconductors. Our findings, along with the mild polarization dependence of the response, provides an explanation for recent THG measurements in cuprates, setting the framework for the theoretical understanding of non-linear effects in unconventional cuprates., Comment: 16 pages, 9 figures

It is an open secret that ImageNet is treated as the panacea of pretraining. Particularly in medical machine learning, models not trained from scratch are often finetuned based on ImageNet-pretrained models. We posit that pretraining on data from the domain of the downstream task should almost always be preferred instead. We leverage RadNet-12M, a dataset containing more than 12 million computed tomography (CT) image slices, to explore the efficacy of self-supervised pretraining on medical and natural images. Our experiments cover intra- and cross-domain transfer scenarios, varying data scales, finetuning vs. linear evaluation, and feature space analysis. We observe that intra-domain transfer compares favorably to cross-domain transfer, achieving comparable or improved performance (0.44% - 2.07% performance increase using RadNet pretraining, depending on the experiment) and demonstrate the existence of a domain boundary-related generalization gap and domain-specific learned features., Comment: Code available from https://github.com/TIO-IKIM/Transfer-learning-across-domain-boundaries/

Nuclei detection and segmentation in hematoxylin and eosin-stained (H&E) tissue images are important clinical tasks and crucial for a wide range of applications. However, it is a challenging task due to nuclei variances in staining and size, overlapping boundaries, and nuclei clustering. While convolutional neural networks have been extensively used for this task, we explore the potential of Transformer-based networks in this domain. Therefore, we introduce a new method for automated instance segmentation of cell nuclei in digitized tissue samples using a deep learning architecture based on Vision Transformer called CellViT. CellViT is trained and evaluated on the PanNuke dataset, which is one of the most challenging nuclei instance segmentation datasets, consisting of nearly 200,000 annotated Nuclei into 5 clinically important classes in 19 tissue types. We demonstrate the superiority of large-scale in-domain and out-of-domain pre-trained Vision Transformers by leveraging the recently published Segment Anything Model and a ViT-encoder pre-trained on 104 million histological image patches - achieving state-of-the-art nuclei detection and instance segmentation performance on the PanNuke dataset with a mean panoptic quality of 0.50 and an F1-detection score of 0.83. The code is publicly available at https://github.com/TIO-IKIM/CellViT, Comment: 18 pages, 5 figures, appendix included

Anisotropic transport properties have been assessed in a number of cuprate superconductors, providing evidence for a nematic state. Here, we analyze recent experimental data for ultrathin YBa$_2$Cu$_3$O$_{7-x}$ (YBCO) films, where nematicity is induced via strain engineering, leading to a suppression of charge density wave scattering along the orthorhombic a-axis and a concomitant enhancement of strange metal behavior along the b-axis. It is shown that the anisotropic properties strongly depend on the substrate, which we characterize by atomic force microscopy (AFM). Based on the AFM data, we provide a microscopic model that can account for the absence (presence) of nematicity and the resulting transport properties in films grown on SrTiO$_3$ (MgO) substrates., Comment: 7 pages, 5 figures

Purpose: Interpreting chest radiographs (CXR) remains challenging due to the ambiguity of overlapping structures such as the lungs, heart, and bones. To address this issue, we propose a novel method for extracting fine-grained anatomical structures in CXR using pseudo-labeling of three-dimensional computed tomography (CT) scans. Methods: We created a large-scale dataset of 10,021 thoracic CTs with 157 labels and applied an ensemble of 3D anatomy segmentation models to extract anatomical pseudo-labels. These labels were projected onto a two-dimensional plane, similar to the CXR, allowing the training of detailed semantic segmentation models for CXR without any manual annotation effort. Results: Our resulting segmentation models demonstrated remarkable performance on CXR, with a high average model-annotator agreement between two radiologists with mIoU scores of 0.93 and 0.85 for frontal and lateral anatomy, while inter-annotator agreement remained at 0.95 and 0.83 mIoU. Our anatomical segmentations allowed for the accurate extraction of relevant explainable medical features such as the cardio-thoracic-ratio. Conclusion: Our method of volumetric pseudo-labeling paired with CT projection offers a promising approach for detailed anatomical segmentation of CXR with a high agreement with human annotators. This technique may have important clinical implications, particularly in the analysis of various thoracic pathologies., Comment: 28 pages, 1 table, 10 figures