Ground Truth Research Articles

Deformable motion compensation in interventional cone-beam CT with a context-aware learned autofocus metric.

Interventional Cone-Beam CT (CBCT) offers 3D visualization of soft-tissue and vascular anatomy, enabling 3D guidance of abdominal interventions. However, its long acquisition time makes CBCT susceptible to patient motion. Image-based autofocus offers a suitable platform for compensation of deformable motion in CBCT, but it relies on handcrafted motion metrics based on first-order image properties and that lack awareness of the underlying anatomy. This work proposes a data-driven approach to motion quantification via a learned, context-aware, deformable metric, , that quantifies the amount of motion degradation as well as the realism of the structural anatomical content in the image. The proposed was modeled as a deep convolutional neural network (CNN) trained to recreate a reference-based structural similarity metric-visual information fidelity (VIF). The deep CNN acted on motion-corrupted images, providing an estimation of the spatial VIF map that would be obtained against a motion-free reference, capturing motion distortion, and anatomic plausibility. The deep CNN featured a multi-branch architecture with a high-resolution branch for estimation of voxel-wise VIF on a small volume of interest. A second contextual, low-resolution branch provided features associated to anatomical context for disentanglement of motion effects and anatomical appearance. The deep CNN was trained on paired motion-free and motion-corrupted data obtained with a high-fidelity forward projection model for a protocol involving 120kV and 9.90mGy. The performance of was evaluated via metrics of correlation with ground truth and with the underlying deformable motion field in simulated data with deformable motion fields with amplitude ranging from 5 to 20mm and frequency from 2.4 up to 4 cycles/scan. Robustness to variation in tissue contrast and noise levels was assessed in simulation studies with varying beam energy (90-120kV) and dose (1.19-39.59mGy). Further validation was obtained on experimental studies with a deformable phantom. Final validation was obtained via integration of on an autofocus compensation framework, applied to motion compensation on experimental datasets and evaluated via metric of spatial resolution on soft-tissue boundaries and sharpness of contrast-enhanced vascularity. The magnitude and spatial map of showed consistent and high correlation levels with the ground truth in both simulation and real data, yielding average normalized cross correlation (NCC) values of 0.95 and 0.88, respectively. Similarly, achieved good correlation values with the underlying motion field, with average NCC of 0.90. In experimental phantom studies, properly reflects the change in motion amplitudes and frequencies: voxel-wise averaging of the local across the full reconstructed volume yielded an average value of 0.69 for the case with mild motion (2mm, 12 cycles/scan) and 0.29 for the case with severe motion (12mm, 6 cycles/scan). Autofocus motion compensation using resulted in noticeable mitigation of motion artifacts and improved spatial resolution of soft tissue and high-contrast structures, resulting in reduction of edge spread function width of 8.78% and 9.20%, respectively. Motion compensation also increased the conspicuity of contrast-enhanced vascularity, reflected in an increase of 9.64% in vessel sharpness. The proposed , featuring a novel context-aware architecture, demonstrated its capacity as a reference-free surrogate of structural similarity to quantify motion-induced degradation of image quality and anatomical plausibility of image content. The validation studies showed robust performance across motion patterns, x-ray techniques, and anatomical instances. The proposed anatomy- and context-aware metric poses a powerful alternative to conventional motion estimation metrics, and a step forward for application of deep autofocus motion compensation for guidance in clinical interventional procedures.

Improved 3D human face reconstruction from 2D images using blended hard edges

AbstractThis study reports an effective and robust edge-based scheme for the reconstruction of 3D human faces from input of single images, addressing drawbacks of existing methods in case of large face pose angles or noisy input images. Accurate 3D face reconstruction from 2D images is important, as it can enable a wide range of applications, such as face recognition, animations, games and AR/VR systems. Edge features extracted from 2D images contain wealthy and robust 3D geometric information, which were used together with landmarks for face reconstruction purpose. However, the accurate reconstruction of 3D faces from contour features is a challenging task, since traditional edge or contour detection algorithms introduce a great deal of noise, which would adversely affect the reconstruction. This paper reports on the use of a hard-blended face contour feature from a neural network and a Canny edge extractor for face reconstruction. The quantitative results indicate that our method achieves a notable improvement in face reconstruction with a Euclidean distance error of 1.64 mm and a normal vector distance error of 1.27 mm when compared to the ground truth, outperforming both traditional and other deep learning-based methods. These metrics show particularly significant advancements, especially in face shape reconstruction under large pose angles. The method also achieved higher accuracy and robustness on in-the-wild images under conditions of blurring, makeup, occlusion and poor illumination.

Comparing novel smartphone pose estimation frameworks with the Kinect V2 for knee tracking during athletic stress tests.

To compare the accuracy of the Microsoft Kinect V2 with novel pose estimation frameworks, in assessing knee kinematics during athletic stress tests, for fast and portable risk assessment of anterior cruciate ligament (ACL) injury. We captured 254 varsity athletes, using the Kinect V2 and a smartphone application utilizing Google's MediaPipe framework. The devices were placed as close as possible and used to capture a person, facing the cameras, performing one of three athletic stress tests at a distance of 2.5ms. Custom software translated the results from both frameworks to the same format. We then extracted relevant knee angles at key moments of the jump and compared them, using the Kinect V2 as the ground truth. The results show relatively small angle differences between the two solutions in the coronal plane and moderate angle differences on the sagittal plane. Overall, the MediaPipe framework results seem to underestimate both knee valgus angles and knee sagittal angles compared to the Kinect V2. This preliminary study demonstrates the potential for Google's MediaPipe framework to be used for calculating lower limb kinematics during athletic stress test motions, which can run on most modern smartphones, as it produces similar results to the Kinect V2. A smartphone application similar to the one developed could potentially be used for low cost and widespread ACL injury prevention.

Alternating low-rank tensor reconstruction for improved multiparametric mapping with cardiovascular MR Multitasking.

To develop a novel low-rank tensor reconstruction approach leveraging the complete acquired data set to improve precision and repeatability of multiparametric mapping within the cardiovascular MR Multitasking framework. A novel approach that alternated between estimation of temporal components and spatial components using the entire data set acquired (i.e., including navigator data and imaging data) was developed to improve reconstruction. The precision and repeatability of the proposed approach were evaluated on numerical simulations, 10 healthy subjects, and 10 cardiomyopathy patients at multiple scan times for 2D myocardial T1/T2 mapping with MR Multitasking and were compared with those of the previous navigator-derived fixed-basis approach. In numerical simulations, the proposed approach outperformed the previous fixed-basis approach with lower T1 and T2 error against the ground truth at all scan times studied and showed better motion fidelity. In human subjects, the proposed approach showed no significantly different sharpness or T1/T2 measurement and significantly improved T1 precision by 20%-25%, T2 precision by 10%-15%, T1 repeatability by about 30%, and T2 repeatability by 25%-35% at 90-s and 50-s scan times The proposed approach at the 50-s scan time also showed comparable results with that of the previous fixed-basis approach at the 90-s scan time. The proposed approach improved precision and repeatability for quantitative imaging with MR Multitasking while maintaining comparable motion fidelity, T1/T2 measurement, and septum sharpness and had the potential for further reducing scan time from 90 s to 50 s.

Maps, Apps and Race: The Market as a Theoretical Machine

AbstractMy commentary situates Desiree Fields' lecture within a long‐standing movement that, paying attention to real estate data and techniques of property, turned the housing market into a theoretical machine. I understand her intervention as a fourth and pivotal conceptual moment in the study of housing markets and inequalities for the discipline of geography, for it brings racial capitalism and settler colonialism as theoretical starting points to analyse the political economy of housing in the era of digital property technologies. As such, her lecture opens up a welcome, ambitious, yet also uncertain terrain to advance global urban theory under platform capitalism and stimulate relational approaches in housing studies. Discussing the regularities through which race, technology and property shape housing and wealth inequalities across variegated sites of market‐making provides a conceptual horizontality that acknowledges the United States as a testing ground, rather than a ground truth, for the production of geographical knowledge.

Automated sleep staging on reduced channels in children with epilepsy

ObjectivesThis study aimed to validate a sleep staging algorithm using in-hospital video-electroencephalogram (EEG) in children without epilepsy, with well-controlled epilepsy (WCE), and with drug-resistant epilepsy (DRE).MethodsOvernight video-EEG, along with electrooculogram (EOG) and chin electromyogram (EMG), was recorded in children between 4 and 18 years of age. Classical sleep staging was performed manually as a ground truth. An end-to-end hierarchical recurrent neural network for sequence-to-sequence automatic sleep staging (SeqSleepNet) was used to perform automated sleep staging using three channels: C4-A1, EOG, and chin EMG.ResultsIn 176 children sleep stages were manually scored: 47 children without epilepsy, 74 with WCE, and 55 with DRE. The 5-class sleep staging accuracy of the automatic sleep staging algorithm was 84.7% for the children without epilepsy, 83.5% for those with WCE, and 80.8% for those with DRE (Kappa of 0.79, 0.77, and 0.73 respectively). Performance per sleep stage was assessed with an F1 score of 0.91 for wake, 0.50 for N1, 0.83 for N2, 0.84 for N3, and 0.86 for rapid eye movement (REM) sleep.ConclusionWe concluded that the tested algorithm has a high accuracy in children without epilepsy and with WCE. Performance in children with DRE was acceptable, but significantly lower, which could be explained by a tendency of more time spent in N1, and by abundant interictal epileptiform discharges and intellectual disability leading to less recognizable sleep stages. REM sleep time, however, significantly affected in children with DRE, can be detected reliably by the algorithm.Clinical trial registration: ClinicalTrials.gov, identifier NCT04584385.

Comprehensive Analysis of Landform and Land Use/ Land Cover Mapping of Waghora Watershed Using Sentinal-2B and SRTM-DEM Data

This paper aims to integrate the use of high spectral resolution a remote sensing data of Sentinel-2B imagery along with the Digital Elevation Model (DEM) derived from the SRTM stereo data to delineate the landforms and analyze the land use and land cover in Waghora watershed, Sausar tehsil, Chhindwara district, situated on basaltic terrain in northern Deccan plateau. A 30 m resolution DEM of the study area was generated, capturing terrain parameters like elevation, slope, aspect, hillshade, and drainage. Sentinel-2B (10 m) images from two seasons were collected FCC (false colour composite) and were prepared by ARCGIS software. combination of Sentinel-2B imagery, SRTM-DEM data, and ground truth verifications to delineate eight major landforms in the studied area. Slope information, extracted from high-resolution DEMs, reveals 4 distinct slope classes. The landforms, including Plateau, Escarpment, Foot Slope, Upper Pediment, Lower Pediment, Valley, Interfluve, and Mound, are characterized based on visual interpretation and geological context. Each landform exhibits unique features, elevation ranges, and dominant land use/cover types, such as single crop cultivation, forest, or degraded land. This detailed terrain analysis aids in land-use planning and environmental management initiatives. The land use/land cover analysis utilizes Sentinel-2B data to identify five classes: degraded forest, double crop, single crop, orchard, and land with and without scrubs. Dense forest dominates 24.2% of the watershed, while double crop and single crop lands cover 36.7% and 24.2% respectively. Orchards occupy a minimal 0.8% of the area, and wastelands account for 5.7%. The relationship between landforms, slope, and land use/cover reveals dominant forestation on steep escarpments, while valleys and interfluves support double cropping due to gentle slopes and depositional processesThe study elucidated connections to examine diverse biophysical phenomena. It highlights the effectiveness of employing SRTM-DEM and Sentinel-2B imagery for analyzing geomorphic characteristics and mapping land use/land cover, offering significant perspectives for land management and planning approaches.

Structured Generation Method of 3D Synthetic Tree Models for Precision Assessment

Abstract. The technology for 3D reconstruction of tree models based on point clouds has been extensively researched, necessitating effective datasets for the study of branch and leaf separation, skeleton point extraction, and tree parameter extraction methods. However, existing datasets for 3D tree models face several challenges, including insufficient data volume for deep learning network training, low accuracy of model ground truth impeding effective method precision evaluation, and a lack of dataset richness to satisfy the needs of multi-type method assessments. In response to these challenges, This paper introduces, for the first time, a fully automated method for generating structured three-dimensional synthetic tree models, and constructs a large-scale 3D synthetic tree dataset enriched with comprehensive structural information. This method facilitates automated computation across several processes, including the mass generation of simulated trees, separation of branches and leaves, noise generation, extraction of skeleton points, and volume calculation. To validate the usability of this dataset across various applications, this paper employs state-of-the-art (SoTA) algorithms to verify the accuracy of methods in 3D tree model reconstruction and carbon stock calculation, thereby thoroughly demonstrating the dataset’s effectiveness.

Experiments with Combining LiDAR and Camera Data Acquired by UAS and Smartphone to Support Mapping

Abstract. Aerial mapping using Unmanned Aerial Systems (UAS), such as the DJI Mavic 3 Enterprise, has revolutionized photogrammetry, enabling efficient data capture for small-scale projects. The typical nadir perspective of UAS mapping, however, imposes limitations on capturing critical details of features due to its predominantly vertical viewpoint. Overcoming this challenge often requires manual, low-altitude flights by experienced UAS pilots to achieve high-angle oblique perspectives, unless gimbled camera mount is used. This study explores the integration of high oblique angle perspectives using the iPhone 15 Pro, which boasts advanced camera capabilities and an integrated LiDAR sensor, to complement UAS imagery. The iPhone 15 Pro's camera sensors provide a Ground Sampled Distance (GSD) comparable to UAS cameras, while its LiDAR sensor, with about five meters of range, enhances mapping capabilities by delivering accurate depth measurements in close range. By utilizing various georeferencing options for the imagery and LiDAR data from the iPhone 15 Pro with UAS nadir imagery, we can achieve a more comprehensive object space reconstruction, significantly improving the accuracy of geospatial mapping. Both the Mavic 3 Enterprise and the iPhone 15 Pro, though operating independently on their respective platforms, support Real-Time Kinematic (RTK) corrections, facilitating precise positioning for the entire system trajectory. Strategic placement and utilization of Ground Control Points (GCPs) aid in the georeferencing of the complete dataset, enhancing its overall accuracy. To validate the accuracy of the acquired data, checkpoints are established on-site. The positions derived from the integrated UAS and iPhone 15 Pro data are compared against these checkpoints to quantify the accuracy and reliability of the data. Additionally, Post-Processed Kinematic (PPK) techniques are employed to validate the trajectories of all data collection systems, ensuring the reliability of the acquired data, especially in instances where RTK corrections may be lacking. In summary, this research showcases comprehensive, multi-dimensional geospatial datasets by conducting validation studies that assess the accuracy and reliability of georeferenced datasets against known ground truth checkpoints. Such validation studies are crucial for identifying gaps in current methodologies and suggesting areas for improvement, thereby aiming to enhance the quality and accuracy of geospatial mapping applications. Through the integration of UAS and smartphone mapping, complemented by rigorous validation efforts, we aspire to achieve improved geospatial mapping accuracy.

A simulation study: comparing independent component analysis and signal-space projection – source-informed reconstruction for rejecting muscle artifacts evoked by transcranial magnetic stimulation

IntroductionThe combination of transcranial magnetic stimulation (TMS) and electroencephalography (EEG) allows researchers to explore cortico-cortical connections. To study effective connections, the first few tens of milliseconds of the TMS-evoked potentials are the most critical. Yet, TMS-evoked artifacts complicate the interpretation of early-latency data. Data-processing strategies like independent component analysis (ICA) and the combined signal-space projection–source-informed reconstruction approach (SSP–SIR) are designed to mitigate artifacts, but their objective assessment is challenging because the true neuronal EEG responses under large-amplitude artifacts are generally unknown. Through simulations, we quantified how the spatiotemporal properties of the artifacts affect the cleaning performances of ICA and SSP–SIR.MethodsWe simulated TMS-induced muscle artifacts and superposed them on pre-processed TMS–EEG data, serving as the ground truth. The simulated muscle artifacts were varied both in terms of their topography and temporal profiles. The signals were then cleaned using ICA and SSP–SIR, and subsequent comparisons were made with the ground truth data.ResultsICA performed better when the artifact time courses were highly variable across the trials, whereas the effectiveness of SSP–SIR depended on the congruence between the artifact and neuronal topographies, with the performance of SSP–SIR being better when difference between topographies was larger. Overall, SSP–SIR performed better than ICA across the tested conditions. Based on these simulations, SSP–SIR appears to be more effective in suppressing TMS-evoked muscle artifacts. These artifacts are shown to be highly time-locked to the TMS pulse and manifest in topographies that differ substantially from the patterns of neuronal potentials.DiscussionSelecting between ICA and SSP–SIR should be guided by the characteristics of the artifacts. SSP–SIR might be better equipped for suppressing time-locked artifacts, provided that their topographies are sufficiently different from the neuronal potential patterns of interest, and that the SSP–SIR algorithm can successfully find those artifact topographies from the high-pass-filtered data. ICA remains a powerful tool for rejecting artifacts that are not strongly time locked to the TMS pulse.

Hierarchical Vector-Quantized Variational Autoencoder and Vector Credibility Mechanism for High-Quality Image Inpainting

Image inpainting infers the missing areas of a corrupted image according to the information of the undamaged part. Many existing image inpainting methods can generate plausible inpainted results from damaged images with the fast-developed deep-learning technology. However, they still suffer from over-smoothed textures or textural distortion in the cases of complex textural details or large damaged areas. To restore textures at a fine-grained level, we propose an image inpainting method based on a hierarchical VQ-VAE with a vector credibility mechanism. It first trains the hierarchical VQ-VAE with ground truth images to update two codebooks and to obtain two corresponding vector collections containing information on ground truth images. The two vector collections are fed to a decoder to generate the corresponding high-fidelity outputs. An encoder then is trained with the corresponding damaged image. It generates vector collections approximating the ground truth by the help of the prior knowledge provided by the codebooks. After that, the two vector collections pass through the decoder from the hierarchical VQ-VAE to produce the inpainted results. In addition, we apply a vector credibility mechanism to promote vector collections from damaged images and approximate vector collections from ground truth images. To further improve the inpainting result, we apply a refinement network, which uses residual blocks with different dilation rates to acquire both global information and local textural details. Extensive experiments conducted on several datasets demonstrate that our method outperforms the state-of-the-art ones.

Training and Comparison of nnU-Net and DeepMedic Methods for Autosegmentation of Pediatric Brain Tumors.

Tumor segmentation is essential in surgical and treatment planning and response assessment and monitoring in pediatric brain tumors, the leading cause of cancer-related death among children. However, manual segmentation is time-consuming and has high interoperator variability, underscoring the need for more efficient methods. After training, we compared 2 deep-learning-based 3D segmentation models, DeepMedic and nnU-Net, with pediatric-specific multi-institutional brain tumor data based on multiparametric MR images. Multiparametric preoperative MR imaging scans of 339 pediatric patients (n = 293 internal and n = 46 external cohorts) with a variety of tumor subtypes were preprocessed and manually segmented into 4 tumor subregions, ie, enhancing tumor, nonenhancing tumor, cystic components, and peritumoral edema. After training, performances of the 2 models on internal and external test sets were evaluated with reference to ground truth manual segmentations. Additionally, concordance was assessed by comparing the volume of the subregions as a percentage of the whole tumor between model predictions and ground truth segmentations using the Pearson or Spearman correlation coefficients and the Bland-Altman method. The mean Dice score for nnU-Net internal test set was 0.9 (SD, 0.07) (median, 0.94) for whole tumor; 0.77 (SD, 0.29) for enhancing tumor; 0.66 (SD, 0.32) for nonenhancing tumor; 0.71 (SD, 0.33) for cystic components, and 0.71 (SD, 0.40) for peritumoral edema, respectively. For DeepMedic, the mean Dice scores were 0.82 (SD, 0.16) for whole tumor; 0.66 (SD, 0.32) for enhancing tumor; 0.48 (SD, 0.27) for nonenhancing tumor; 0.48 (SD, 0.36) for cystic components, and 0.19 (SD, 0.33) for peritumoral edema, respectively. Dice scores were significantly higher for nnU-Net (P ≤ .01). Correlation coefficients for tumor subregion percentage volumes were higher (0.98 versus 0.91 for enhancing tumor, 0.97 versus 0.75 for nonenhancing tumor, 0.98 versus 0.80 for cystic components, 0.95 versus 0.33 for peritumoral edema in the internal test set). Bland-Altman plots were better for nnU-Net compared with DeepMedic. External validation of the trained nnU-Net model on the multi-institutional Brain Tumor Segmentation Challenge in Pediatrics (BraTS-PEDs) 2023 data set revealed high generalization capability in the segmentation of whole tumor, tumor core (a combination of enhancing tumor, nonenhancing tumor, and cystic components), and enhancing tumor with mean Dice scores of 0.87 (SD, 0.13) (median, 0.91), 0.83 (SD, 0.18) (median, 0.89), and 0.48 (SD, 0.38) (median, 0.58), respectively. The pediatric-specific data-trained nnU-Net model is superior to DeepMedic for whole tumor and subregion segmentation of pediatric brain tumors.

Tomographic reconstructions of the fast-ion phase space using imaging neutral particle analyser measurements

In this paper we demonstrate how the inversion, in energy and major radius (E, R) coordinates, of imaging neutral particle analyser (INPA) measurements can be used to obtain the fast-ion distribution. The INPA is most sensitive to passing ions with energies in the range (20–150) keV and pitches near 0.5 in the core and 0.7 near the plasma edge. Inversion of synthetic signals, via 0th-order Tikhonov and Elastic Net regularization, were performed to demonstrate the capability of recovering the ground truth fast-ion 2D phase-space distribution resolved in major radius and energy, even in the presence of moderate noise levels (10%). Finally, we apply our method to measure the 2D phase-space distribution in an MHD quiescent plasma at ASDEX Upgrade and find good agreement with the slowing down fast-ion distribution predicted by TRANSP.

Impact of deep learning image reconstruction on volumetric accuracy and image quality of pulmonary nodules with different morphologies in low-dose CT

BackgroundThis study systematically compares the impact of innovative deep learning image reconstruction (DLIR, TrueFidelity) to conventionally used iterative reconstruction (IR) on nodule volumetry and subjective image quality (IQ) at highly reduced radiation doses. This is essential in the context of low-dose CT lung cancer screening where accurate volumetry and characterization of pulmonary nodules in repeated CT scanning are indispensable.Materials and methodsA standardized CT dataset was established using an anthropomorphic chest phantom (Lungman, Kyoto Kaguku Inc., Kyoto, Japan) containing a set of 3D-printed lung nodules including six diameters (4 to 9 mm) and three morphology classes (lobular, spiculated, smooth), with an established ground truth. Images were acquired at varying radiation doses (6.04, 3.03, 1.54, 0.77, 0.41 and 0.20 mGy) and reconstructed with combinations of reconstruction kernels (soft and hard kernel) and reconstruction algorithms (ASIR-V and DLIR at low, medium and high strength). Semi-automatic volumetry measurements and subjective image quality scores recorded by five radiologists were analyzed with multiple linear regression and mixed-effect ordinal logistic regression models.ResultsVolumetric errors of nodules imaged with DLIR are up to 50% lower compared to ASIR-V, especially at radiation doses below 1 mGy and when reconstructed with a hard kernel. Also, across all nodule diameters and morphologies, volumetric errors are commonly lower with DLIR. Furthermore, DLIR renders higher subjective IQ, especially at the sub-mGy doses. Radiologists were up to nine times more likely to score the highest IQ-score to these images compared to those reconstructed with ASIR-V. Lung nodules with irregular margins and small diameters also had an increased likelihood (up to five times more likely) to be ascribed the best IQ scores when reconstructed with DLIR.ConclusionWe observed that DLIR performs as good as or even outperforms conventionally used reconstruction algorithms in terms of volumetric accuracy and subjective IQ of nodules in an anthropomorphic chest phantom. As such, DLIR potentially allows to lower the radiation dose to participants of lung cancer screening without compromising accurate measurement and characterization of lung nodules.

ECG-image-kit: a synthetic image generation toolbox to facilitate deep learning-based electrocardiogram digitization

Abstract Objective: Cardiovascular diseases are a major cause of mortality globally, and electrocardiograms (ECGs) are crucial for diagnosing them. Traditionally, ECGs are stored in printed formats. However, these printouts, even when scanned, are incompatible with advanced ECG diagnosis software that require time-series data. Digitizing ECG images is vital for training machine learning models in ECG diagnosis, leveraging the extensive global archives collected over decades. Deep learning models for image processing are promising in this regard, although the lack of clinical ECG archives with reference time-series data is challenging. Data augmentation techniques using realistic generative data models provide a solution.

Approach: We introduce ECG-Image-Kit, an open-source toolbox for generating synthetic multi-lead ECG images with realistic artifacts from time-series data, aimed at automating the conversion of scanned ECG images to ECG data points. The tool synthesizes ECG images from real time-series data, applying distortions like text artifacts, wrinkles, and creases on a standard ECG paper background.

Main results: As a case study, we used ECG-Image-Kit to create a dataset of 21,801 ECG images from the PhysioNet QT database. We developed and trained a combination of a traditional computer vision and deep neural network model on this dataset to convert synthetic images into time-series data for evaluation. We assessed digitization quality by calculating the signal-to-noise ratio (SNR) and compared clinical parameters like QRS width, RR, and QT intervals recovered from this pipeline, with the ground truth extracted from ECG time-series. The results show that this deep learning pipeline accurately digitizes paper ECGs, maintaining clinical parameters, and highlights a generative approach to digitization.

Significance: The toolbox has broad applications, including model development for ECG image digitization and classification. The toolbox currently supports data augmentation for the 2024 PhysioNet Challenge, focusing on digitizing and classifying paper ECG images.

Interpretable (not just posthoc-explainable) medical claims modeling for discharge placement to reduce preventable all-cause readmissions or death

We developed an inherently interpretable multilevel Bayesian framework for representing variation in regression coefficients that mimics the piecewise linearity of ReLU-activated deep neural networks. We used the framework to formulate a survival model for using medical claims to predict hospital readmission and death that focuses on discharge placement, adjusting for confounding in estimating causal local average treatment effects. We trained the model on a 5% sample of Medicare beneficiaries from 2008 and 2011, based on their 2009–2011 inpatient episodes (approximately 1.2 million), and then tested the model on 2012 episodes (approximately 400 thousand). The model scored an out-of-sample AUROC of approximately 0.75 on predicting all-cause readmissions—defined using official Centers for Medicare and Medicaid Services (CMS) methodology—or death within 30-days of discharge, being competitive against XGBoost and a Bayesian deep neural network, demonstrating that one need-not sacrifice interpretability for accuracy. Crucially, as a regression model, it provides what blackboxes cannot—its exact gold-standard global interpretation, explicitly defining how the model performs its internal “reasoning” for mapping the input data features to predictions. In doing so, we identify relative risk factors and quantify the effect of discharge placement. We also show that the posthoc explainer SHAP provides explanations that are inconsistent with the ground truth model reasoning that our model readily admits.

What Fechner could not do: Separating perceptual encoding and decoding with difference scaling.

A key question in perception research is how stimulus variations translate into perceptual magnitudes, that is, the perceptual encoding process. As experimenters, we cannot probe perceptual magnitudes directly, but infer the encoding process from responses obtained in a psychophysical experiment. The most prominent experimental technique to measure perceptual appearance is matching, where observers adjust a probe stimulus to match a target in its appearance along the dimension of interest. The resulting data quantify the perceived magnitude of the target in physical units of the probe, and are thus an indirect expression of the underlying encoding process. In this paper, we show analytically and in simulation that data from matching tasks do not sufficiently constrain perceptual encoding functions, because there exist an infinite number of pairs of encoding functions that generate the same matching data. We use simulation to demonstrate that maximum likelihood conjoint measurement (Ho, Landy, & Maloney, 2008; Knoblauch & Maloney, 2012) does an excellent job of recovering the shape of ground truth encoding functions from data that were generated with these very functions. Finally, we measure perceptual scales and matching data for White's effect (White, 1979) and show that the matching data can be predicted from the estimated encoding functions, down to individual differences.

Mapping surface sediment characteristics in enclosed shallow‐marine environments using spatially balanced designs and the random forest algorithm

AbstractMapping the sedimentary character of the seafloor in large water‐filled basins is fundamental for understanding landform dynamics to inform research, management, intervention and conservation actions. Seabed mapping methods have undergone considerable development in the last two decades, including the uptake of machine learning approaches for sediment size prediction and classification. However, predictions of surficial sediment characteristics are often hindered by the availability of ground truthing data, their arrangement in space and the modelling approach chosen. Spatially informed sampling designs provide an opportunity to significantly improve the accuracy and uncertainty of predicted sediment distributions. In this study, we apply a machine learning algorithm to predict sediment distributions across Port Phillip Bay, a large (1930 km2) structurally controlled estuary on the southeast coast of Victoria, Australia. Surface sediment samples (n = 252) were collected using a spatially balanced design, ensuring that sampling effort was spread evenly within the embayment with increased sampling intensity placed in more heterogeneous areas. Surficial textural metrics were modelled using the random forest algorithm with bathymetric and hydrodynamic predictor variables. Models highlighted trends in sediment grain size, sorting and composition consistent with predicted wave‐ and current‐induced sediment mobilisation. Model predictions were accurate (normalised‐root mean squared error [NRMSE]: 0.14–0.16); however, standard error was not homogeneous across the study area. Uncertainty maps highlighted areas where additional sampling effort may be needed, including areas where transitional bathymetry impacted surficial sediment character and areas of anthropogenic modifications to the seabed. This study shows the benefits of undertaking spatially informed sample design, block cross‐validation during model fitting and quantifying spatial uncertainty in predictive maps to accurately quantify the fundamental boundary conditions of sediment size. The results of this study are intended to inform local coastal management, including beach renourishment activities. However, approaches outlined are applicable to any study where the seafloor grain size is a fundamental variable in understanding landscape change.

Cross Domain Early Crop Mapping with Label Spaces Discrepancies using MultiCropGAN

Abstract. Mapping target crops before the harvest season for regions lacking crop-specific ground truth is critical for global food security. Utilizing multispectral remote sensing and domain adaptation methods, prior studies strive to produce precise crop maps in these regions (target domain) with the help of the crop-specific labelled remote sensing data from the source regions (source domain). However, existing approaches assume identical label spaces across those domains, a challenge often unmet in reality, necessitating a more adaptable solution. This paper introduces the Multiple Crop Mapping Generative Adversarial Neural Network (MultiCropGAN) model, comprising a generator, discriminator, and classifier. The generator transforms target domain data into the source domain, employing identity losses to retain the characteristics of the target data. The discriminator aims to distinguish them and shares the structure and weights with the classifier, which locates crops in the target domain using the generator’s output. This model’s novel capability lies in locating target crops within the target domain, overcoming differences in crop type label spaces between the target and source domains. In experiments, MultiCropGAN is benchmarked against various baseline methods. Notably, when facing differing label spaces, MultiCropGAN significantly outperforms other baseline methods. The Overall Accuracy is improved by about 10%.

Use of Artificial Intelligence With Deep Learning Approaches for the Follow-up of Infrarenal Endovascular Aortic Repair.

Endoleaks represent one of the main complications after endovascular aortic repair (EVAR) and can lead to increased re-intervention rates and secondary rupture. Serial lifelong surveillance is required and traditionally involves cross-sectional imaging with manual axial measurements. Artificial intelligence (AI)-based imaging analysis has been developed and may provide a more precise and faster assessment. This study aims to evaluate the ability of an AI-based software to assess post-EVAR morphological changes over time, detect endoleaks, and associate them with EVAR-related adverse events. Patients who underwent EVAR at a tertiary hospital from January 2017 to March 2020 with at least 2 follow-up computed tomography angiography (CTA) were analyzed using PRAEVAorta 2 (Nurea). The software was compared to the ground truth provided by human experts using Sensitivity (Se), Specificity (Sp), Negative Predictive Value (NPV), and Positive Predictive Value (PPV). Endovascular aortic repair-related adverse events were defined as aneurysm-related death, rupture, endoleak, limb occlusion, and EVAR-related re-interventions. Fifty-six patients were included with a median imaging follow-up of 27 months (interquartile range [IQR]: 20-40). There were no significant differences overtime in the evolution of maximum aneurysm diameters (55.62 mm [IQR: 52.33-59.25] vs 54.34 mm [IQR: 46.13-59.47]; p=0.2162) or volumes (130.4 cm3 [IQR: 113.8-171.7] vs 125.4 cm3 [IQR: 96.3-169.1]; p=0.1131) despite a -13.47% decrease in the volume of thrombus (p=0.0216). PRAEVAorta achieved a Se of 89.47% (95% confidence interval [CI]: 80.58 to 94.57), a Sp of 91.25% (95% CI: 83.02 to 95.70), a PPV of 90.67% (95% CI: 81.97 to 95.41), and an NPV of 90.12% (95% CI: 81.70 to 94.91) in detecting endoleaks. Endovascular aortic repair-related adverse events were associated with global volume modifications with an area under the curve (AUC) of 0.7806 vs 0.7277 for maximum diameter. The same trend was observed for endoleaks (AUC of 0.7086 vs 0.6711). The AI-based software PRAEVAorta enabled a detailed anatomic characterization of aortic remodeling post-EVAR and showed its potential interest for automatic detection of endoleaks during follow-up. The association of aortic aneurysmal volume with EVAR-related adverse events and endoleaks was more robust compared with maximum diameter. The integration of PRAEVAorta AI software into clinical practice promises a transformative shift in post-EVAR surveillance. By offering precise and rapid detection of endoleaks and comprehensive anatomic assessments, clinicians can expect enhanced diagnostic accuracy and streamlined patient management. This innovation reduces reliance on manual measurements, potentially reducing interpretation errors and shortening evaluation times. Ultimately, PRAEVAorta's capabilities hold the potential to optimize patient care, leading to more timely interventions and improved outcomes in endovascular aortic repair.