Search

We introduce VascX models, a comprehensive set of model ensembles for analyzing retinal vasculature from color fundus images (CFIs). Annotated CFIs were aggregated from public datasets . Additional CFIs, mainly from the population-based Rotterdam Study were annotated by graders for arteries and veins at pixel level, resulting in a dataset diverse in patient demographics and imaging conditions. VascX models demonstrated superior segmentation performance across datasets, image quality levels, and anatomic regions when compared to existing, publicly available models, likely due to the increased size and variety of our training set. Important improvements were observed in artery-vein and disc segmentation performance, particularly in segmentations of these structures on CFIs of intermediate quality, common in large cohorts and clinical datasets. Importantly, these improvements translated into significantly more accurate vascular features when we compared features extracted from VascX segmentation masks with features extracted from segmentation masks generated by previous models. With VascX models we provide a robust, ready-to-use set of model ensembles and inference code aimed at simplifying the implementation and enhancing the quality of automated retinal vasculature analyses. The precise vessel parameters generated by the model can serve as starting points for the identification of disease patterns in and outside of the eye.

Satellite Communications (SatCom) are a backbone of worldwide development. In contrast with the past, when the GEO satellites were the only means for such connectivity, nowadays the multi-orbital connectivity is emerging, especially with the use of satellite constellations. Simultaneously, SatCom enabled the so-called In-Flight Connectivity, while with the advent of 5G-NTN, the development of this market is being accelerated. However, there are still various missing points before such a technology becomes mainstream, especially in the case of Rotary Wing Aircraft (RWA). Indeed, due to their particular characteristics, such as the low altitude flights and the blade interference, there are still open challenges. In this work, an End-to-End (E2E) analysis for the performance of SatCom under 5G-NTN for manned and unmanned RWA is performed. Various scenarios are examined, and related requirements are shown. The effects of blades and other characteristics of the RWA are established, and simulations for these cases are developed. Results along with related discussion are presented, while future directions for development are suggested. This work is part of the ESA ACROSS-AIR project., Comment: Accepted for publication at the 29th IEEE CAMAD Workshop, 21-23 Oct, 2024, Athens, Greece

Largely due to their implicit nature, neural fields lack a direct mechanism for filtering, as Fourier analysis from discrete signal processing is not directly applicable to these representations. Effective filtering of neural fields is critical to enable level-of-detail processing in downstream applications, and support operations that involve sampling the field on regular grids (e.g. marching cubes). Existing methods that attempt to decompose neural fields in the frequency domain either resort to heuristics or require extensive modifications to the neural field architecture. We show that via a simple modification, one can obtain neural fields that are low-pass filtered, and in turn show how this can be exploited to obtain a frequency decomposition of the entire signal. We demonstrate the validity of our technique by investigating level-of-detail reconstruction, and showing how coarser representations can be computed effectively., Comment: Project Page: https://theialab.github.io/banf

We present conjectures on the scattering terms of cluster scattering diagrams of rank 2, supported by significant computational evidence., Comment: v2: Expanded and corrected the introductory material, divided the paper into sections, and updated the section on related work. The statements of the conjectures are unchanged

For affine Coxeter groups of affine types $\tilde D$ and $\tilde B$, we model the interval $[1,c]_T$ in the absolute order by symmetric noncrossing partitions of an annulus with one or two double points. In type $\tilde B$ (and \emph{almost} in type $\tilde D$), the diagrams also model the larger lattice defined by McCammond and Sulway., Comment: 51 pages, 6 figures. Version 2: Small revisions and improvements throughout

We prove a simple formula for arbitrary cluster variables in the marked surfaces model. As part of the formula, we associate a labeled poset to each tagged arc, such that the associated $F$-polynomial is a weighted sum of order ideals. Each element of the poset has a weight, and the weight of an ideal is the product of the weights of the elements of the ideal. In the unpunctured case, the weight on each element is a single $\hat{y}_i$, in the usual sense of principal coefficients. In the presence of punctures, some elements may have weights of the form $\hat{y}_i/\hat{y}_j$. Our search for such a formula was inspired by the Fundamental Theorem of Finite Distributive Lattices combined with work of Gregg Musiker, Ralf Schiffler, and Lauren Williams that, in some cases, organized the terms of the $F$-polynomial into a distributive lattice. The proof consists of a simple and poset-theoretically natural argument in a special case, followed by a hyperbolic geometry argument using a cover of the surface to prove the general case., Comment: 43 pages, 41 figures. Version 2: Corrected a minor error in the .tex file. Version 3: Minor changes to improve references to other papers. Version 4: Minor changes to the introduction

Neural Radiance Fields (NeRFs) have shown promise in applications like view synthesis and depth estimation, but learning from multiview images faces inherent uncertainties. Current methods to quantify them are either heuristic or computationally demanding. We introduce BayesRays, a post-hoc framework to evaluate uncertainty in any pre-trained NeRF without modifying the training process. Our method establishes a volumetric uncertainty field using spatial perturbations and a Bayesian Laplace approximation. We derive our algorithm statistically and show its superior performance in key metrics and applications. Additional results available at: https://bayesrays.github.io.

The field of ice sheet, ice shelf and glacier-related seismology, cryoseismology, has seen rapid development in recent years. As concern grows for the implications of change in the great ice sheets of Greenland and Antarctica, so instrument advances and international field programs have expanded the availability of passive seismic datasets that contain records of glacier processes. Many glacier-related seismic signals are generated by melt water, so a need exists to better understand the seismic response of moving water on glaciers. We present an analytic framework to model seismic signals at the interface between moving water and ice based on either new or existing smoothed particle hydrodynamic (SPH) simulations. The framework is tailored to understanding the seismic body wave response of transient moving water events and produces a simulated dataset of water-ice collisions to predict the motion due to body waves at any location on the glacier. The synthetic acceleration time series generated are invariant to the simulation resolution, and consider frequency-dependent weak dispersion and attenuation due to the propagating medium. We illustrate the capability of our framework using end-member cases of water flow: the breaking of a supraglacial melt water lake dam with the subsequent flow of water through glacier channels with differing geometries. Our focus is on the waveform attributes rather than exact waveform matching in view of the dynamic nature of the glacier environment. The flexibility inherent in the computational framework will allow for the simulation of seismic signals generated by high-energy collisions in a variety of different water flow geometries. We make the code publicly available for the polar geophysics community with the aim of adding to the toolbox of available approaches to inform the future monitoring of melt water movement and related glacier processes., Comment: 21 pages, 9 figures, 0 tables; submitted to GJI

The noncrossing partition poset associated to a Coxeter group $W$ and Coxeter element $c$ is the interval $[1,c]_T$ in the absolute order on $W$. We construct a new model of noncrossing partititions for $W$ of classical affine type, using planar diagrams (affine types $\tilde A$ and $\tilde C$ in this paper and affine types $\tilde D$ and $\tilde B$ in the sequel). The model in type $\tilde A$ consists of noncrossing partitions of an annulus. In type $\tilde C$, the model consists of symmetric noncrossing partitions of an annulus or noncrossing partitions of a disk with two orbifold points. Following the lead of McCammond and Sulway, we complete $[1,c]_T$ to a lattice by factoring the translations in $[1,c]_T$, but the combinatorics of the planar diagrams leads us to make different choices about how to factor., Comment: 46 pages, 12 figures. Version 2: Minor changes in connection with a revision of arXiv:2212.13799. Version 3: Minor changes and corrections in connection with the release of arXiv:2312.17331. Version 4: to appear in Comb. Theory

We define noncrossing partitions of a marked surface without punctures (interior marked points). We show that the natural partial order on noncrossing partitions is a graded lattice and describe its rank function topologically. Lower intervals in the lattice are isomorphic to products of noncrossing partition lattices of other surfaces. We similarly define noncrossing partitions of a symmetric marked surface with double points and prove some of the analogous results. The combination of symmetry and double points plays a role that one might have expected to be played by punctures., Comment: 39 pages, 13 figures. Version 3: Removed the assertion that the noncrossing partitions of a symmetric marked surface with double points form a lattice and provided a counterexample. Version 4: Minor changes, additions, and corrections in connection with the release of arXiv:2312.17331. Version 5: to appear in SIAM J. Discrete Math

The "Effectiveness of Early Literacy Instruction: Summary of 20 Years of Research" study examined information about evidence-based practices that are essential for early childhood educators and policymakers as they support preschool children's language and literacy development. The study used a process modeled after the What Works Clearinghouse (WWC) methodology to systematically identify effective early childhood curricula, lesson packages, instructional practices, and technology programs in studies conducted from 1997 to 2017. The study team identified 132 interventions evaluated by 109 studies that the study team determined were high-quality experimental or quasi-experimental studies. The WWC's evidence standards are used to assess the quality of an evaluation study and the strength of its claims about whether an intervention caused the observed effect on student achievement. This document presents the following six appendixes that accompany the study: (1) The search, screening, and review process; (2) Technical results; (3) Effects of interventions by domain and outcome type for the 132 interventions evaluated by high-quality impact studies; (4) Implementation characteristics of the 132 interventions evaluated by high-quality impact studies; (5) Research basis for the 109 high-quality impact studies; and (6) Three-hundred fifty-seven studies reviewed using the evidence standards. [For the full report, see ED613947. For the Study Snapshot, see ED613950.]

Children entering kindergarten vary greatly in their language and literacy skills. Therefore, up-to-date information about evidence-based practices is essential for early childhood educators and policymakers as they support preschool children's language and literacy development. This study used a process modeled after the What Works Clearinghouse (WWC) methodology to systematically identify effective early childhood curricula, lesson packages, instructional practices, and technology programs in studies conducted from 1997 to 2017. More than 74,000 studies were analyzed to identify interventions that improved students' performance in six language and literacy domains (language, phonological awareness, print knowledge, decoding, early writing, and general literacy). The study team identified 132 interventions evaluated by 109 studies that the study team determined were high-quality experimental or quasi-experimental studies. The WWC's evidence standards are used to assess the quality of an evaluation study and the strength of its claims about whether an intervention caused the observed effect on student achievement. To better understand the effectiveness of the interventions, their implementation characteristics and instructional features were coded for the relevant language and literacy domains. The findings revealed that instruction that teaches a specific domain is likely to increase performance in that domain. Interventions that teach language exclusively might be more beneficial when conducted in small groups or one-on-one than in larger group sizes. In addition, teaching both phonological awareness and print knowledge might benefit performance in print knowledge. Finally, some evidence indicates that instruction that teaches both phonological awareness and print knowledge might also lead to improvements in decoding and early writing performance. [For the Study Snapshot, see ED613950. For the appendixes, see ED613954.]

3D multi-object tracking (MOT) is a key problem for autonomous vehicles, required to perform well-informed motion planning in dynamic environments. Particularly for densely occupied scenes, associating existing tracks to new detections remains challenging as existing systems tend to omit critical contextual information. Our proposed solution, InterTrack, introduces the Interaction Transformer for 3D MOT to generate discriminative object representations for data association. We extract state and shape features for each track and detection, and efficiently aggregate global information via attention. We then perform a learned regression on each track/detection feature pair to estimate affinities, and use a robust two-stage data association and track management approach to produce the final tracks. We validate our approach on the nuScenes 3D MOT benchmark, where we observe significant improvements, particularly on classes with small physical sizes and clustered objects. As of submission, InterTrack ranks 1st in overall AMOTA among methods using CenterPoint detections., Comment: Accepted to CRV 2023

District leaders in a large urban school district in central Florida wanted to examine the efficacy of a new curriculum designed to enhance the word knowledge of grade 5 students so as to improve reading achievement. The new curriculum, called Word Knowledge Instruction (WKI), consists of 15-minute lessons 4 days a week for 20 weeks. The lessons address state standards and cover 20 prefixes and suffixes. Thirty-nine schools participated in the study, with 92 English language arts (ELA) teachers in high-poverty schools randomly assigned within schools either to use WKI or to continue to use their standard ELA curriculum. Classroom observations revealed that WKI was implemented as intended. WKI had a positive effect, equivalent to an increase of 9 percentile points, on students' ability to correctly extract and spell a base word from a derived word, one of the skills explicitly taught by WKI. WKI had no effect on two other related reading skills that were not directly taught by WKI (students' ability to select a nonword that best fits the grammatical context of a sentence or to use knowledge of word parts to infer meaning of new words) or on students' vocabulary or reading scores. These findings suggest that, although students learned what they were explicitly taught, the transferability to related but not directly taught skills might require more intense or longer duration instruction or additional professional development for teachers. [For the study snapshot, see ED611684. For the appendix, see ED611685.]

This is the appendix for the report, "The Impact of Word Knowledge Instruction on Literacy Outcomes in Grade 5." District leaders in a large urban school district in central Florida wanted to examine the efficacy of a new curriculum designed to enhance the word knowledge of grade 5 students so as to improve reading achievement. The new curriculum, called Word Knowledge Instruction (WKI), consists of 15-minute lessons 4 days a week for 20 weeks. The lessons address state standards and cover 20 prefixes and suffixes. Thirty-nine schools participated in the study, with 92 English language arts (ELA) teachers in high-poverty schools randomly assigned within schools either to use WKI or to continue to use their standard ELA curriculum. This appendix describes the study data, intervention, analytic sample, and analyses used to assess the impact of Word Knowledge Instruction (WKI) on student literacy outcomes in grade 5. One appendix is included in this document: Data, intervention, sample, and methodology. [For the full report, see ED611683. For the study snapshot, see ED611684.]

The purpose was two-fold: To explore the capability of the Diamond Synchrotron infra-red so as to include near-field photothermal microspectroscopy (PTMS); and Toward a quantitative nanohistology - investigation of scleroderma using synchrotron radiation (mu- FTIR). With recent advances in AFM, the integration of an IR temperature-based system on an IR beamline is still a promising and unique approach. These preliminary tests of the PTMS stand-alone system with the Diamond Multimode InfraRed Imaging and Microspectroscopy beamline (MIRIAM) were successful in one beamtime. To our knowledge, this is still among very few studies worldwide that managed to interface an AFM-resistive probe to measure an infrared spectrum using synchrotron radiation (SR) as a source. When the alignment was successful, we managed to obtain good spectra on samples that were strong infrared absorbers. However, the alignment could not be guaranteed all the time, and the signal-to-noise ratio became too low to record any meaningful spectra on samples deemed of scientific interest for this proposal. We clarified the possible improvements to PTMS instrumentation and data acquisition needed to ensure that we can collect sufficient thermal signals from very thin samples., Comment: 10 pages.12 figures, two additional references with relevant comments added to text

We give an explicit construction of the cluster scattering diagram for any acyclic exchange matrix of affine type. We show that the corresponding cluster scattering fan coincides both with the mutation fan and with a fan constructed in the almost-positive roots model., Comment: 45 pages, 2 figures. Version 2: Fixed typographical error

These are the appendixes for the report "Effects of an Inquiry-Oriented Curriculum and Professional Development Program on Grade 7 Students' Understanding of Statistics and on Statistics Instruction." This cluster-randomized controlled trial compared the Supporting Teacher Enactment of the Probability and Statistics Standards (STEPSS) program--a 20-day curriculum unit designed to support teaching and learning of the state curriculum standards for statistics and probability in grade 7, along with four days of professional development for teachers--with practice-as-usual probability and statistics instruction and teacher professional development. The two primary outcomes of interest were student understanding of statistics and classroom instruction in statistics. Two appendices are included in the document: (1) About the study; and (2) Study data and data analysis. [For the full report, see ED610166. For the study snapshot, see ED610167.]

On average, Florida students earn only half of the points possible in the statistics content area of the state's annual mathematics assessment. Leaders in Broward County Public Schools, a large, diverse, urban school district, viewed changes to statistics curriculum and instruction as one way to address this issue. This study randomly assigned 40 middle schools in the district to either implement a replacement curriculum unit with four days of teacher professional development in probability and statistics or continue with their practice-as-usual instruction in probability and statistics. The replacement unit supported teaching and learning of all the probability and statistics standards in the grade 7 course description. The replacement unit with the associated professional development, called the Supporting Teacher Enactment of the Probability and Statistics Standards program, improved student understanding of statistics and statistics instruction. The magnitude of the effect on student understanding was 23 percent of 1 standard deviation, which is comparable to an increase of 9 percentile points for an average student. [For the appendixes, see ED610168. For the study snapshot, see ED610167.]

We have implemented an extension for the observational seismology obspy software package to provide a streamlined tool tailored to the processing of seismic signals from non-earthquake sources, in particular those from deforming systems such as glaciers and landslides. This seismic attributes library provides functionality to: (1) download and/or pre-process seismic waveform data; (2) detect and catalogue seismic events using multi-component signals from one or more seismometers; and (3) calculate characteristics ('attributes'/'features') of the identified events. The workflow is controlled by three main functions that have been tested for the breadth of data types expected from permanent and campaign-deployed seismic instrumentation. A selected STA/LTA-type (short-term average/long-term average), or other, event detection algorithm can be applied to the waveforms and user-defined functions implemented to calculate any required characteristics of the detected events. The code is written in Python 2/3 and is available on GitHub together with detailed documentation and worked examples., Comment: 13 pages, 3 figures; submitted to JORS. Code available at https://github.com/rossjturner/seismic\_attributes

Accurate measurement of the local component of geodetic motion at GPS stations presents a challenge due to the need to separate this signal from the tectonic plate rotation. A pressing example is the observation of glacial isostatic adjustment (GIA) which constrains the Earth's response to ice unloading, and hence, contributions of ice-covered regions such as Antarctica to global sea level rise following ice mass loss. We focus on horizontal GPS velocities which typically contain a large component of plate rotation and a smaller local component primarily relating to GIA. Incomplete separation of these components introduces significant bias into estimates of GIA motion vectors. We present the results of a series of tests based on the motions of GPS stations from East Antarctica: 1) signal separation for sets of synthetic data that replicate the geometric character of non-separable, and separable, GIA-like horizontal velocities; and 2) signal separation for real GPS station data with an appraisal of uncertainties. For both synthetic and real motions, we compare results where the stations are unweighted, and where each station is areal-weighted using a metric representing the inverse of the spatial density of neighbouring stations. From the synthetic tests, we show that a GIA-like signal is recoverable from the plate rotation signal providing it has geometric variability across East Antarctica. We also show that areal-weighting has a very significant effect on the ability to recover a GIA-like signal with geometric variability, and hence on separating the plate rotation and local components. For the real data, assuming a rigid Antarctic plate, fitted plate rotation parameters compare well with other studies in the literature. We find that 25 out of 36 GPS stations examined in East Antarctica have non-zero local horizontal velocities, at the 2$\sigma$ level, after signal separation., Comment: 15 pages, 9 figures, 5 tables; published in GJI

The Southern Ocean (in the region 60-180$^\circ$E) south of the Indian Ocean, Australia, and the West Pacific is noted for the frequent occurrence and severity of its storms. These storms give rise to high-amplitude secondary microseisms from sources, including the deep ocean regions, and primary microseisms where the swells impinge on submarine topographic features. A better understanding of the varying microseism wavefield enables improvements to seismic imaging and development of proxy observables to complement sparse in situ wave observations and hindcast models of the global ocean wave climate. We analyze 12-26 years of seismic data from 11 seismic stations either on the East Antarctic coast or sited in the Indian Ocean, Australia, and New Zealand. The power spectral density of the seismic wavefield is calculated to explore how the time-changing microseism intensity varies with (i) sea ice coverage surrounding Antarctica and (ii) the Southern Annular Mode (SAM) climate index. Variations in sea ice extent are found to be the dominant control on the microseism intensity at Antarctic stations, which exhibit a seasonal pattern phase-shifted by 4-5 months compared to stations in other continents. Peaks in extremal intensity at East Antarctic stations occur in March-April, with the highest peaks for secondary microseisms occurring during negative SAM events. This relationship between microseism intensity and the SAM index is opposite to that observed on the Antarctic Peninsula. This work informs the complexity of microseism amplitudes in the Southern Hemisphere and assists ongoing interdisciplinary investigations of interannual variability and long-term trends., Comment: 16 pages, 8 figures, 3 tables; published in JGR Earth Surface

An instantaneous sub-surface disturbance in a two-dimensional elastic half-space is considered. The disturbance propagates through the elastic material until it reaches the free surface, after which it propagates out along the surface. In conventional theory, the free-surface conditions on the unknown surface are projected onto the flat plane $y = 0$, so that a linear model may be used. Here, however, we present a formulation that takes explicit account of the fact that the location of the free surface is unknown {\it a priori}, and we show how to solve this more difficult problem numerically. This reveals that, while conventional linearized theory gives an accurate account of the decaying waves that travel outwards along the surface, it can under-estimate the strength of the elastic rebound above the location of the disturbance. In some circumstances, the non-linear solution fails in finite time, due to the formation of a ``peakon'' at the free surface. We suggest that brittle failure of the elastic material might in practice be initiated at those times and locations., Comment: 26 pages, 7 figures

Monocular 3D object detection is a key problem for autonomous vehicles, as it provides a solution with simple configuration compared to typical multi-sensor systems. The main challenge in monocular 3D detection lies in accurately predicting object depth, which must be inferred from object and scene cues due to the lack of direct range measurement. Many methods attempt to directly estimate depth to assist in 3D detection, but show limited performance as a result of depth inaccuracy. Our proposed solution, Categorical Depth Distribution Network (CaDDN), uses a predicted categorical depth distribution for each pixel to project rich contextual feature information to the appropriate depth interval in 3D space. We then use the computationally efficient bird's-eye-view projection and single-stage detector to produce the final output bounding boxes. We design CaDDN as a fully differentiable end-to-end approach for joint depth estimation and object detection. We validate our approach on the KITTI 3D object detection benchmark, where we rank 1st among published monocular methods. We also provide the first monocular 3D detection results on the newly released Waymo Open Dataset. We provide a code release for CaDDN which is made available., Comment: Accepted in CVPR 2021

This paper presents two approaches to mathematical modelling of a synthetic seismic pulse, and a comparison between them. First, a new analytical model is developed in two-dimensional Cartesian coordinates. Combined with an initial condition of sufficient symmetry, this provides a valuable check for the validity of the numerical method that follows. A particular initial condition is found which allows for a new closed-form solution. A numerical scheme is then presented which combines a spectral (Fourier) representation for displacement components and wave-speed parameters, a fourth order Runge-Kutta integration method, and an absorbing boundary layer. The resulting large system of differential equations is solved in parallel on suitable enhanced performance desktop hardware in a new software implementation. This provides an alternative approach to forward modelling of waves within isotropic media which is efficient, and tailored to rapid and flexible developments in modelling seismic structure, for example, shallow depth environmental applications. Visual comparisons of the analytic solution and the numerical scheme are presented., Comment: 18 pages, 6 figures