Ground Truth Research Articles

Shape completion in the dark: completing vertebrae morphology from 3D ultrasound.

Ultrasound (US) imaging, while advantageous for its radiation-free nature, is challenging to interpret due to only partially visible organs and a lack of complete 3D information. While performing US-based diagnosis or investigation, medical professionals therefore create a mental map of the 3D anatomy. In this work, we aim to replicate this process and enhance the visual representation of anatomical structures. We introduce a point cloud-based probabilistic deep learning (DL) method to complete occluded anatomical structures through 3D shape completion and choose US-based spine examinations as our application. To enable training, we generate synthetic 3D representations of partially occluded spinal views by mimicking US physics and accounting for inherent artifacts. The proposed model performs consistently on synthetic and patient data, with mean and median differences of 2.02 and 0.03 in Chamfer Distance (CD), respectively. Our ablation study demonstrates the importance of US physics-based data generation, reflected in the large mean and median difference of 11.8 CD and 9.55 CD, respectively. Additionally, we demonstrate that anatomical landmarks, such as the spinous process (with reconstruction CD of 4.73) and the facet joints (mean distance to ground truth (GT) of 4.96 mm), are preserved in the 3D completion. Our work establishes the feasibility of 3D shape completion for lumbar vertebrae, ensuring the preservation of level-wise characteristics and successful generalization from synthetic to real data. The incorporation of US physics contributes to more accurate patient data completions. Notably, our method preserves essential anatomical landmarks and reconstructs crucial injections sites at their correct locations.

Cardiac MRI segmentation using shifted-window multilayer perceptron mixer networks.

In this work, we proposed a deep-learning segmentation algorithm for cardiac magnetic resonance imaging (MRI) to aid in contouring of the left ventricle (LV), right ventricle (RV), and Myocardium (Myo). 
Methods: We proposed a shifted window multilayer perceptron (Swin-MLP) mixer network which is built upon a 3D U-shaped symmetric encoder-decoder structure. We evaluated our proposed network using public data from 100 individuals. The network performance was quantitatively evaluated using 3D volume similarity between the ground truth contours and the predictions using Dice score coefficient, sensitivity, and precision as well as 2D surface similarity using Hausdorff distance (HD), mean surface distance (MSD) and residual mean square distance (RMSD). We benchmarked the performance against two other current leading edge networks known as Dynamic UNet and Swin-UNetr on the same public dataset. 
Results: The proposed network achieved the following volume similarity metrics when averaged over three cardiac segments: Dice = 0.952±0.017, precision = 0.948±0.016, sensitivity = 0.956±0.022. The average surface similarities were HD = 1.521±0.121 mm, MSD = 0.266±0.075 mm, and RMSD = 0.668±0.288 mm. The network shows statistically significant improvement in comparison to the Dynamic UNet and Swin-UNetr algorithms for most volumetric and surface metrics with p-value less than 0.05. Overall, the proposed Swin-MLP mixer network demonstrates better or comparable performance than competing methods. 
Conclusions: The proposed Swin-MLP mixer network demonstrates more accurate segmentation performance compared to current leading edge methods. This robust method demonstrates the potential to streamline clinical workflows for multiple applications.

Self-Supervised 3D Scene Flow Estimation and Motion Prediction using Local Rigidity Prior.

In this article, we investigate self-supervised 3D scene flow estimation and class-agnostic motion prediction on point clouds. A realistic scene can be well modeled as a collection of rigidly moving parts, therefore its scene flow can be represented as a combination of rigid motion of these individual parts. Building upon this observation, we propose to generate pseudo scene flow labels for self-supervised learning through piecewise rigid motion estimation, in which the source point cloud is decomposed into local regions and each region is treated as rigid. By rigidly aligning each region with its potential counterpart in the target point cloud, we obtain a region-specific rigid transformation to generate its pseudo flow labels. To mitigate the impact of potential outliers on label generation, when solving the rigid registration for each region, we alternately perform three steps: establishing point correspondences, measuring the confidence for the correspondences, and updating the rigid transformation based on the correspondences and their confidence. As a result, confident correspondences will dominate label generation, and a validity mask will be derived for the generated pseudo labels. By using the pseudo labels together with their validity mask for supervision, models can be trained in a self-supervised manner. Extensive experiments on FlyingThings3D and KITTI datasets demonstrate that our method achieves new state-of-the-art performance in self-supervised scene flow learning, without any ground truth scene flow for supervision, even performing better than some supervised counterparts. Additionally, our method is further extended to class-agnostic motion prediction and significantly outperforms previous state-of-the-art self-supervised methods on nuScenes dataset.

GRouNdGAN: GRN-guided simulation of single-cell RNA-seq data using causal generative adversarial networks

We introduce GRouNdGAN, a gene regulatory network (GRN)-guided reference-based causal implicit generative model for simulating single-cell RNA-seq data, in silico perturbation experiments, and benchmarking GRN inference methods. Through the imposition of a user-defined GRN in its architecture, GRouNdGAN simulates steady-state and transient-state single-cell datasets where genes are causally expressed under the control of their regulating transcription factors (TFs). Training on six experimental reference datasets, we show that our model captures non-linear TF-gene dependencies and preserves gene identities, cell trajectories, pseudo-time ordering, and technical and biological noise, with no user manipulation and only implicit parameterization. GRouNdGAN can synthesize cells under new conditions to perform in silico TF knockout experiments. Benchmarking various GRN inference algorithms reveals that GRouNdGAN effectively bridges the existing gap between simulated and biological data benchmarks of GRN inference algorithms, providing gold standard ground truth GRNs and realistic cells corresponding to the biological system of interest.

Semi-TSGAN: Semi-Supervised Learning for Highlight Removal Based on Teacher-Student Generative Adversarial Network

Despite recent notable advancements in highlight image restoration techniques, the dearth of annotated data and the lightweight deployment of highlight removal networks pose significant impediments to further advancements in the field. In this paper, to the best of our knowledge, we first propose a semi-supervised learning paradigm for highlight removal, merging the fusion version of a teacher–student model and a generative adversarial network, featuring a lightweight network architecture. Initially, we establish a dependable repository to house optimal predictions as pseudo ground truth through empirical analyses guided by the most reliable No-Reference Image Quality Assessment (NR-IQA) method. This method serves to assess rigorously the quality of model predictions. Subsequently, addressing concerns regarding confirmation bias, we integrate contrastive regularization into the framework to curtail the risk of overfitting on inaccurate labels. Finally, we introduce a comprehensive feature aggregation module and an extensive attention mechanism within the generative network, considering a balance between network performance and computational efficiency. Our experimental evaluations encompass comprehensive assessments on both full-reference and non-reference highlight benchmarks. The results demonstrate conclusively the substantive quantitative and qualitative enhancements achieved by our proposed algorithm in comparison to state-of-the-art methodologies.

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.

AgriTera: Accurate Non-Invasive Fruit Ripeness Sensing via Sub-Terahertz Wireless Signals

The potential to transform the agriculture sector and reduce food waste significantly lies in the ability to assess the quality of fruits and vegetables on a large scale. This article introduces AgriTera, an innovative solution that utilizes sub-terahertz wireless signals for accurate and non-invasive fruit ripeness sensing, without the need for physical contact. The underlying idea is that sugar and water concentrations in fruits, linked to ripening, create distinctive patterns in the wideband spectrum of reflected signals from the fruit body. AgriTera leverages the sub-THz bands due to their wide bandwidth, sensitivity to water, millimeter-scale penetration depth, and non-ionizing features. This enables high-resolution inferences from both the peel and pulp underneath. A chemometric model is developed to translate reflection spectra into well-known ripeness metrics like Dry Matter and Brix. Extensive over-the-air experiments using commercial sub-THz transceivers are conducted and compared with ground truth values from specialized sensors and a vision-based scheme.