Equivalent Performance Research Articles

Process parameter optimization in polymer powder bed fusion of final part properties in polyphenylene sulfide through design of experiments

AbstractThe Additive Manufacturing (AM) modality of Laser-Based Powder Bed Fusion of Polymers (PBF-LB/P) is an established method for manufacturing semi-crystalline polymers. Like other AM processes, the selection of PBF-LB/P process parameters is critical as it has direct effect on final part properties. While prior research has been predominantly focused on polyamides (e.g., nylon 12), there exists a gap in exploring how process parameters affect higher performance polymers, such as polyphenylene sulfide (PPS). This work aims to explore the effects of PBF-LB/P process parameters on PPS parts printed via PBF-LB/P. While prior PBF-LB/P parameter research primarily relies on evaluating energy input to the system through a single numerical value of energy density, this study investigates the interplay of the print parameters within the energy density equation. To achieve these goals, an analysis was performed on the influence of the laser power, hatch spacing, and beam velocity on ultimate tensile strength (UTS), modulus, and crystallinity of printed parts. A Taguchi L8 array was used in balancing the print parameter combinations allowing for isolation of variance to the specific factors and interactions. Through this approach, print parameter combinations that improved UTS and modulus were identified. Additionally, the study revealed that numerically equivalent energy densities did not lead to equivalent performance, underscoring the significance for including the constitutive process parameters within the energy density equation when establishing process property relationships in printing with PBF-LB/P.

Equivalent combined cycle modeling and performance optimization for a three-heat-reservoir thermal Brownian heat transformer with external heat-transfer

Three-heat-reservoir (THR) heat transformers can upgrade temperature gradient of thermal energy, and lots of instructive research has been conducted in the last decades. However, study of THR heat transformer cycle theory in micro systems remains lacking. By employing macro equivalent combined cycle method, this paper builds a finite-time thermodynamic model of THR thermal Brownian heat transformer, which is a combination of a two-reservoir thermal Brownian heat pump driven by a two-reservoir thermal Brownian engine. Expressions of performance parameters are deduced, and operating temperatures and external load ratio are determined by solving heat balance equations. Maximal heating load and corresponding coefficient of performance (COP) are given by modulating external load, heat exchanger inventory allocations and barrier height synchronously, and the optimal thermal conductance allocations and optimal working temperatures are identified. Results indicate that external thermal resistances affect heat flow transmission and the coupling of combined cycle, ultimately shaping the cycle performance. Half of the overall heat exchanger inventory should be arranged in middle heat exchanger under maximal heating load objective. Heating load exhibits an extremum with respect to COP. Equivalent combined cycle modelling is an effective and efficient method for performance optimization of THR thermal Brownian heat transformers with external heat-transfer.

Universal representations in cardiovascular ECG assessment: A self-supervised learning approach

BackgroundThe 12-lead electrocardiogram (ECG) is an established modality for cardiovascular assessment. While deep learning algorithms have shown promising results for analyzing ECG data, the limited availability of labeled datasets hinders broader applications. Self-supervised learning can learn meaningful representations from the unlabeled data and transfer the knowledge to downstream tasks. This study underscores the development and validation of a self-supervised learning methodology tailored to produce universal ECG representations from longitudinally collected ECG data, applicable across a spectrum of cardiovascular assessments. MethodsWe introduced a pre-trained model that utilizes contrastive self-supervised learning to universal ECG representations from 4,932,573 ECG tracing from 1,684,298 adult patients on 7 campuses of Chang Gung Memorial Hospital. We extensively evaluated the proposed model using an internal dataset collected from diverse healthcare establishments and an external public dataset encompassing varied cardiovascular conditions and sample magnitudes. ResultsThe pre-trained model showed the equivalent performance to the conventionally trained models, which solely rely on supervised learning in both internal and external datasets, to assess atrial fibrillation, atrial flutter, premature rhythm abnormalities, first-degree atrioventricular block, and myocardial infarction. When applied to small sample sizes, it was observed that the learned ECG representations enhanced the classification models, resulting in an improvement of up to 0.3 of the area under the receiver operating characteristic (AUROC). ConclusionsThe ECG representations learned from longitudinal ECG data are highly effective, particularly with small sample sizes, and further enhance the learning process and boost robustness.

Theory-guided machine learning for thermal modeling of in-situ automated fiber placement of thermoplastic composites

In-situ Automated Fiber Placement (AFP) of thermoplastic composites has several advantages over traditional manufacturing techniques, with the main benefit being eliminating secondary thermal processing. Without secondary heat treatment, the in-situ thermal history becomes the critical process parameter that governs bond development, crystallization kinetics, and the development of residual stresses. This work improves the thermal modeling of the in-situ Automated Fiber Placement (AFP) manufacturing process by leveraging Theory-Guided Machine Learning (TGML). A novel theory-guided neural network (TgNN) with theory-based pre-layer transforms models the three-dimensional temperature distribution during in-situ AFP manufacturing. The TgNN is fit on experimentally measured temperatures for various combinations of hot gas torch temperatures and heat source velocities. Feature engineering is implemented by applying theory-based pre-layer transforms to the input features time, the thermocouple coordinates, hot gas torch temperature, and heat source velocity. Compared to a theory-agnostic neural network, the TgNN with theory-based pre-layer transforms has improved predictive ability and requires fewer training data for equivalent performance. The trained model is computationally efficient and can be leveraged for online process control.

Comprehensive Symptom Prediction in Inpatients With Acute Psychiatric Disorders Using Wearable-Based Deep Learning Models: Development and Validation Study.

Assessing the complex and multifaceted symptoms of patients with acute psychiatric disorders proves to be significantly challenging for clinicians. Moreover, the staff in acute psychiatric wards face high work intensity and risk of burnout, yet research on the introduction of digital technologies in this field remains limited. The combination of continuous and objective wearable sensor data acquired from patients with deep learning techniques holds the potential to overcome the limitations of traditional psychiatric assessments and support clinical decision-making. This study aimed to develop and validate wearable-based deep learning models to comprehensively predict patient symptoms across various acute psychiatric wards in South Korea. Participants diagnosed with schizophrenia and mood disorders were recruited from 4 wards across 3 hospitals and prospectively observed using wrist-worn wearable devices during their admission period. Trained raters conducted periodic clinical assessments using the Brief Psychiatric Rating Scale, Hamilton Anxiety Rating Scale, Montgomery-Asberg Depression Rating Scale, and Young Mania Rating Scale. Wearable devices collected patients' heart rate, accelerometer, and location data. Deep learning models were developed to predict psychiatric symptoms using 2 distinct approaches: single symptoms individually (Single) and multiple symptoms simultaneously via multitask learning (Multi). These models further addressed 2 problems: within-subject relative changes (Deterioration) and between-subject absolute severity (Score). Four configurations were consequently developed for each scale: Single-Deterioration, Single-Score, Multi-Deterioration, and Multi-Score. Data of participants recruited before May 1, 2024, underwent cross-validation, and the resulting fine-tuned models were then externally validated using data from the remaining participants. Of the 244 enrolled participants, 191 (78.3%; 3954 person-days) were included in the final analysis after applying the exclusion criteria. The demographic and clinical characteristics of participants, as well as the distribution of sensor data, showed considerable variations across wards and hospitals. Data of 139 participants were used for cross-validation, while data of 52 participants were used for external validation. The Single-Deterioration and Multi-Deterioration models achieved similar overall accuracy values of 0.75 in cross-validation and 0.73 in external validation. The Single-Score and Multi-Score models attained overall R² values of 0.78 and 0.83 in cross-validation and 0.66 and 0.74 in external validation, respectively, with the Multi-Score model demonstrating superior performance. Deep learning models based on wearable sensor data effectively classified symptom deterioration and predicted symptom severity in participants in acute psychiatric wards. Despite lower computational costs, Multi models demonstrated equivalent or superior performance than Single models, suggesting that multitask learning is a promising approach for comprehensive symptom prediction. However, significant variations were observed across wards, which presents a key challenge for developing clinical decision support systems in acute psychiatric wards. Future studies may benefit from recurring local validation or federated learning to address generalizability issues.

NIMG-76. INCORPORATING IMAGING FEATURES OF NORMAL BRAIN TO ENHANCE SPATIALLY DIFFERENTIATING TREATMENT-INDUCED EFFECTS FROM RECURRENT GLIOMA WITH AI

Abstract INTRODUCTION Although AI-based approaches have been applied to discriminate between tumor recurrence (rTumor) and treatment-induced effects (TxE) from chemoradiotherapy in patients with recurrent glioma, these models typically either do not account for within-lesion mixtures of rTumor and TxE or generalize to independent test-sets with >80% accuracy. We hypothesize that incorporating imaging features of normal-appearing brain (NAB) into existing AI-based models that leverage tissue-samples with known locations on MRI and physiological imaging will significantly improve performance for distinguishing regions of TxE from rTumor and allow for visualization of rTumor beyond the T2-lesion. METHODS Using 254 pathology-confirmed tissue-samples with coordinates on diffusion-weighted, perfusion-weighted, and anatomical MRI from 135 glioma suspected of progression, we modified previously-developed binary, AI-based prediction models to learn multi-class targets consisting of rTumor, TxE, and contralateral NAB-regions (3/patient) with imaging signatures of CSF, white-matter, and grey-matter (659 total 10mm-isotropic samples). The multi-class ensemble model was trained with 4 unique repeats of constrained-5-fold-cross-validation and evaluated on a held-out test set (30 patients). This process was repeated 20 times, and the performance was compared to corresponding binary AI-based models. RESULTS NAB was easiest to distinguish from individual and combined classes (AUC-ROCs>0.98). Multi-class model performance for predicting TxE increased by 27% to 0.93±0.03 (p<0.00001) when discriminating between NAB and rTumor as compared to just rTumor with the binary models. While TxE vs rTumor alone was most difficult to predict with the multi-class model (AUC-ROC=0.73) achieving equivalent performance to binary models, there was 67% less variance in test performance. This approach allowed for the quantification of whole-brain probability maps of rTumor, TxE, and NAB along with a combined 3-channel colormap, facilitating the visualization of rTumor predictions beyond the T2-lesion. CONCLUSIONS Including NAB imaging features improved performance of AI-based models of rTumor/TxE, while expanding the predicted area to produce more interpretable whole-brain probability maps. Current work is evaluating their utility in prospectively guiding sampling of rTumor tissue and predicting survival.

Carangiform‐Like Magnetic Milliswimmer With Negative Buoyancy for Agile 3D Navigation in Confined Fluid Environments

AbstractMiniature soft robots designed with magnetoelastic materials have the potential for noninvasive navigation in narrow spaces and innovative clinical therapies. However, the complex environment inside the body imposes stringent requirements on the motility and environmental adaptability of robots. Inspired by carangiform fish morphology and kinematics, a 3D free‐swimming magnetic milliswimmer with enhanced controllability, maneuverability, and environmental adaptability is developed. This milliswimmer is negatively buoyant like most fishes and employs magnetic torque to actuate its body. It mimics fish muscle contractions and generates a carangiform‐like body curvature distribution that results in comparable swimming behavior and performance. Compared with previous designs, it achieves over six times the equivalent motion performance and enables gravity‐resisting free‐swimming, with fish‐like maneuverability (a minimum turning radius of just 0.05 body length and a maximum turning rate of up to 4737 deg s−1). Its ability to transport and release cargo precisely while navigating three‐dimensionally in an ex vivo porcine urinary system is demonstrated. The designed bionic magnetic soft robot is expected to advance biomedical applications of magnetoelastic materials, particularly in urinary and other clinical treatments.

Use of a Deuterium-Deuterium Neutron Generator for the Add-A-Source Waste Matrix Correction Technique

Neutron generators offer an alternative to radioactive sources in the application of active techniques for nondestructive assay of nuclear materials. The Add-A-Source technique has enabled improvements in the ability to accurately measure waste bearing nuclear materials by measuring and correcting for the non-nuclear matrix effect. The technique has historically been performed with a Cf-252 spontaneous fission source. In this work we have demonstrated the use of a Deuterium-deuterium neutron generator in place of the Cf-252 source. A DD generator enables operational advantages and allows the simplification of the instrument design, while offering equivalent corrective performance as a Cf-252 source.

To develop a classification system which helps differentiate cystic intraductal papillary neoplasm of the bile duct from mucinous cystic neoplasm of the liver

ObjectiveTo establish a classification system which differentiates cystic intraductal papillary neoplasm of the bile duct (cystic IPNB) from hepatic mucinous cystic tumors (MCN) based on their radiological difference. MethodsA total of 75 patients pathologically diagnosed as MCN and IPNB in two major hospitals from 2015 to 2024 were enrolled. Radiological features were recorded and compared between these two tumors. Variables with significant differences were included in multivariate logistic regression (LR) analysis. A decision model was built and simplified based on importance ranking of variables. K-nearest-neighbor (KNN) model was introduced to learn distribution of individuals in main dimensions based on multiple correspondence analysis (MCA) and predicted diagnosis. The diagnostic efficacy of the classification system and the KNN model was compared. ResultsSignificant differences existed in Dmax-IVC angle, septation, mural nodule, upstream and downstream biliary dilatation, communication with bile duct between MCN and cystic IPNB. Downstream biliary dilatation and communication with bile duct were highly specific for IPNB (specificity, 97.9 % and 100 %, respectively), which could independently diagnose IPNB. Among four significant indicators in LR analysis, upstream biliary dilatation and Dmax-IVC angle were used for a simplified decision model to attain good applicability. The KNN model based on MCA data achieved highest accuracy (0.910) when K = 11. Overall, the classification system achieved an AUC of 0.882 (0.95CI: 0.797–0.966), compared with 0.911 (0.95CI: 0.818–1.000) in the KNN model, which demonstrated no significant difference (p = 0.655) in differential performance. ConclusionThe classification system combining four important indicators had equivalent performance to KNN model in discrimination, which was simple and applicable for clinical practice, and also accessible on unenhanced examinations.

Innovative Approaches to Beam Forming Antenna Array Systems with Adaptive Partial Update NLMS Algorithms

Improvements in interference cancellation, energy economy, spectrum efficiency, and system security are among the most pressing needs for today's wireless networks. One effective technique for this is the use of a Beamforming Array Antennas (BAA). However, the complexity of the Beamforming (BF) network, the long convergence time, and the large number of adjustable weight coefficients, all work against full band BAA. The key innovation and contribution of this research was to use Partial Update (PU) instead of full band adaptive algorithms, as no previous attempt had been made to do so. A subset of the array's elements, rather than all of them, will be connected to by PU methods. This allows the system to reduce the number of active antennas across all cells while maintaining high efficiency, and low cost. In this research, a new architectural model was proposed that makes use of PU adaptive algorithms, to reduce the required number of phase shifters (PSs), and hence base station antennas. For the most part, we will be discussing PU Normalized Least Mean Square (PU NLMS) algorithms like M-max NLMS, Periodic-NLMS, and Stochastic-NLMS. Using a Uniform Linear Array (ULA) Antennas in a simulation environment, we find that the, in terms of Mean Square Error (MSE), convergence rate, and steady-state error, it is evident that all PU NLMS algorithms (with the exception of Periodic-NLMS) had performed close, and approximately equivalent performance to the full band NLMS algorithms. In other hands, these PU algorithms maintain the radiation pattern with as little change from the original (distortion-free) as possible and symmetrical of the array, while. fewer elements are needed to produce the same amount of radiation. Reducing the number of required coefficients N to M (M = 5; M: Number of coefficients to be update per iteration) compared to the full update method (N = 8), to obtain the Reduction ratio 38%.

Euclid preparation

Euclid will collect an enormous amount of data during the mission’s lifetime, observing billions of galaxies in the extragalactic sky. Along with traditional template-fitting methods, numerous machine learning (ML) algorithms have been presented for computing their photometric redshifts and physical parameters (PPs), requiring significantly less computing effort while producing equivalent performance measures. However, their performance is limited by the quality and amount of input information entering the model (the features), to a level where the recovery of some well-established physical relationships between parameters might not be guaranteed – for example, the star-forming main sequence (SFMS). To forecast the reliability of Euclid photo-zs and PPs calculations, we produced two mock catalogs simulating the photometry with the UNIONS ugriz and Euclid filters. We simulated the Euclid Wide Survey (EWS) and Euclid Deep Fields (EDF), alongside two auxiliary fields. We tested the performance of a template-fitting algorithm (Phosphoros) and four ML methods in recovering photo-zs, PPs (stellar masses and star formation rates), and the SFMS on the simulated Euclid fields. To mimic the Euclid processing as closely as possible, the models were trained with Phosphoros-recovered labels and tested on the simulated ground truth. For the EWS, we found that the best results are achieved with a mixed labels approach, training the models with wide survey features and labels from the Phosphoros results on deeper photometry, that is, with the best possible set of labels for a given photometry. This imposes a prior to the input features, helping the models to better discern cases in degenerate regions of feature space, that is, when galaxies have similar magnitudes and colors but different redshifts and PPs, with performance metrics even better than those found with Phosphoros. We found no more than 3% performance degradation using a COSMOS-like reference sample or removing u band data, which will not be available until after data release DR1. The best results are obtained for the EDF, with appropriate recovery of photo-z, PPs, and the SFMS.

Investigating the Limits of Familiarity-Based Navigation.

Insect-inspired navigation strategies have the potential to unlock robotic navigation in power-constrained scenarios, as they can function effectively with limited computational resources. One such strategy, familiarity-based navigation, has successfully navigated a robot along routes of up to 60 m using a single-layer neural network trained with an Infomax learning rule. Given the small size of the network that effectively encodes the route, here we investigate the limits of this method, challenging it to navigate longer routes, investigating the relationship between performance, view acquisition rate and dimension, network size, and robustness to noise. Our goal is both to determine the parameters at which this method operates effectively and to explore the profile with which it fails, both to inform theories of insect navigation and to improve robotic deployments. We show that effective memorization of familiar views is possible for longer routes than previously attempted, but that this length decreases for reduced input view dimensions. We also show that the ideal view acquisition rate must be increased with route length for consistent performance. We further demonstrate that computational and memory savings may be made with equivalent performance by reducing the network size-an important consideration for applicability to small, lower-power robots-and investigate the profile of memory failure, demonstrating increased confusion across the route as it extends in length. In this extension to previous work, we also investigate the form taken by the network weights as training extends and the areas of the image on which visual familiarity-based navigation most relies. Additionally, we investigate the robustness of familiarity-based navigation to view variation caused by noise.

Accurate equivalent analysis models and performance evaluation of precast concrete shear walls

Precast concrete shear walls with horizontal steel connections (PCSWHs) were proposed and experimentally studied. To improve the design and analysis efficiency of the PCSWHs with adequate accuracy and reliability, the constitutive models of the connections joints and precast shear wall modules were established using theoretical derivation, and accurate equivalent analysis (AEA) models of the PCSWHs were developed. Furthermore, performance evaluation of three-story and three-span precast frame-shear wall structures with four types of shear walls, including the CAST model (cast-in-situ shear walls), PRE-H model (PCSWHs were adopted), PRE-HVR model (precast concrete shear walls having “strong horizontal and weak vertical” connections, which abbreviated as PCSWHVs, were adopted, and frame beam was connected), and PRE-HVS model (PCSWHVs were adopted, and frame beam can slide freely), was carried out based on the AEA models. Results show that the interaction between the stud of the steel connector and concrete can be simplified as a four-point skeleton curve. The AEA models of the PCSWHs were validated. PCSWHVs can significantly reduce the shear force in the deep beams around the shear walls, and it effectively avoids the undesirable brittle shear failure of deep beams. Compared with the CAST model, the maximum decrease rate of shear force in the deep beams of the PRE-HVS models is 34.56 %. PCSWHVs can fully utilize the plastic energy dissipation of steel shear keys (SSKs). The SSKs in the PRE-HVR and PRE-HVS models enter a plastic yield state under rare earthquakes, and the latter develops greater hysteretic energy dissipation than the former.

360° CZT-SPECT/CT cameras: 99mTc- and 177Lu-phantom-based evaluation under clinical conditions

PurposeFor the first time, three currently available 360° CZT-SPECT/CT cameras were compared under clinical conditions using phantom-based measurements.MethodsA 99mTc- and a 177Lu-customized NEMA IEC body phantom were imaged with three different cameras, StarGuide (GE Healthcare), VERITON-CT versions 200 (V200) and 400 (V400) (Spectrum Dynamics Medical) under the same clinical conditions. Energy resolution and volumetric sensitivity were evaluated from energy spectra. Vendors provided the best reconstruction parameters dedicated to visualization and/or quantification, based on their respective software developments. For both 99mTc- and 177Lu-phantoms, noise level, quantification accuracy, and recovery coefficient (RC) were performed with 3DSlicer. Image quality metrics from an approach called “task-based” were computed with iQMetrix-CT on 99mTc visual reconstructions to assess, through spatial frequencies, noise texture in the background (NPS) and contrast restitution of a hot insert (TTF). Spatial resolution indices were calculated from frequencies corresponding to TTF10% and TTF50%.ResultsDespite the higher sensitivity of VERITON cameras and the enhanced energy resolution of the V400 (3.2% at 140 keV, 5.2% at 113 keV, and 3.6% at 208 keV), StarGuide presents comparable image quality. This highlights the need to differentiate sensitivity from count quality, which is influenced by hardware design (collimator, detector block) and conditions image quality as well as the reconstruction process (algorithms, scatter correction, noise regulation). For 99mTc imaging, the quantitative image optimization approach based on RCmean for StarGuide versus RCmax for V200 and V400 systems (RCmean/RCmax: 0.9/1.8; 0.5/0.9; 0.5/0.9 respectively—Ø37 mm). SRTB10/50 showed nearly equivalent spatial resolution performances across the different reconstructed images. For 177Lu imaging, the 113 keV imaging of the V200 and V400 systems demonstrated strong performances in both image quality and quantification, while StarGuide and V400 systems offer even better potential due to their ability to exploit signals from both the 113 and 208 keV peaks. 177Lu quantification was optimized according to RCmax for all cameras and reconstructions (1.07 ± 0.09—Ø37 mm).ConclusionsThe three cameras have equivalent potential for 99mTc imaging, while StarGuide and V400 have demonstrated higher potential for 177Lu. Dedicated visual or quantitative reconstructions offer better specific performances compared to the unified visual/quantitative reconstruction. The task-based approach appears to be promising for in-depth comparison of images in the context of system characterization/comparison and protocol optimization.

Development of a mathematical model for harbor maneuvers to realize modeling automation

A simulation environment of harbor maneuvers is critical for developing automatic berthing. Mathematical models are widely used to estimate harbor maneuvers. However, user’s analysis and decision are necessary to derive, select, and identify the model because each actuator configuration needs an inherent mathematical expression. We proposed a new mathematical model for arbitrary configurations to overcome that issue. The new model is a hybrid model that combines the simplicity of the derivation of the Taylor expansion and the high degree of freedom of the MMG low-speed maneuvering model. We also developed a method to select mathematical expressions for the proposed model using system identification. Because the proposed model can easily derive mathematical expressions, we can generate multiple expressions simultaneously and choose the best one. This method can reduce the workload of model identification and selection. Furthermore, the proposed method will enable the automatic generation of mathematical models because it can reduce user’s decision-making and data analysis for the model generation due to its less dependency on the knowledge of ship hydrodynamics and captive model test. The proposed method was validated with free-running model tests and showed equivalent or better estimation performance than the conventional model generation method.

Variable Splitting and Fusing for Image Phase Retrieval.

Phase Retrieval is defined as the recovery of a signal when only the intensity of its Fourier Transform is known. It is a non-linear and non-convex optimization problem with a multitude of applications including X-ray crystallography, microscopy and blind deconvolution. In this study, we address the problem of Phase Retrieval from the perspective of variable splitting and alternating minimization for real signals and seek to develop algorithms with improved convergence properties. An exploration of the underlying geometric relations led to the conceptualization of an algorithmic step aiming to refine the estimate at each iteration via recombination of the separated variables. Following this, a theoretical analysis to study the convergence properties of the proposed method and justify the inclusion of the recombination step was developed. Our experiments showed that the proposed method converges substantially faster compared to other state-of-the-art analytical methods while demonstrating equivalent or superior performance in terms of quality of reconstruction and ability to converge under various setups.

EFFECT OF BLENDED LEARNING EDUCATIONAL MODEL ON SECONDARY SCHOOL STUDENTS’ MATHEMATICS CONCEPTUAL UNDERSTANDING

The study was conducted to assess the effect of a blended learning educational model on grade ten students' mathematics conceptual understanding in Ethiopia. A non-equivalent pre-test and post-test quasi-experimental design was employed. Two intact classes from different schools were randomly selected and assigned to experimental (n=38) and comparison group (n=45). From the previous semester, students’ record book revealed that both groups have almost an equivalent mathematics performance. A one-way ANOVA and paired sample t-test were used to analyze the data. Moreover, to check the equivalence of the groups, the same pre-test was developed and administered. The result indicates that both groups were equivalent. The test covers tenth-grade topics relations and functions, polynomial functions, exponential functions, and logarithmic functions. The study shows that students instructed through the blended learning approach surpassed those taught via the conventional method in their comprehension of mathematical concepts. The one-way ANOVA statistical result revealed there is a notable disparity between groups with effect size of η2 =.269, which is a large effect size. Moreover, a paired sample t-test shows the experimental group showed a substantial mean difference of 14.327 with a very large effect size (d=1.016) when compared with the comparison group. Thus, the study shows that blended learning was more effective in improving mathematical conceptual understanding when compared with a conventional method of teaching. Based on the findings, the researcher recommends blended learning in mathematics education and suggests stakeholders and policymakers incorporate the model into the curriculum. Keywords: blended learning, conceptual understanding, educational model, lab rotation model, mathematical concept

Diagnostic Performance for Detection of Glaucomatous Structural Damage Using Pixelwise Analysis of Retinal Thickness Measurements.

To compare the diagnostic accuracy of thickness measurements of individual and combined macular retinal layers to discriminate 188 glaucomatous and 148 glaucoma suspect eyes from 362 healthy control (HC) eyes on a pixel-by-pixel basis. For this retrospective study, we manually corrected the segmentations of posterior pole optical coherence tomography (OCT) scans to determine the thickness of the nerve fiber layer (NFL), ganglion cell layer (GCL), inner plexiform layer (IPL), the ganglion cell complex (GCC), and the total neural retina (TR). For each eye, the total number of pixels with thickness values less than the fifth percentile of the HC distribution was used to create a receiver operating characteristic (ROC) curve for each layer and for layer combinations. Using total abnormal pixel count criteria to discriminate glaucoma from HC eyes, the individual layers with the highest area under the ROC curve (AUC) were the NFL and GCL; IPL performance was significantly lower (P < 0.05). GCC had a significant higher AUC (94.3%) than individual the AUC of the NFL (92.3%) (P = 0.0231) but not higher than AUC of the GCL (93.4%) (P = 0.3487). The highest AUC (95.4%) and sensitivity (85.1%) at 95% specificity was found for the Boolean combination of NFL or GCL. The highest AUC is not significantly higher (P = 0.0882) than the AUC of the GCC but the highest sensitivity is significantly higher than the sensitivity of the GCC. This pattern was similar for discriminating between suspect and HC eyes (P = 0.0356). Using pixel-based methods, the diagnostic accuracy of NFL and GCL exceeded that of IPL and TR. GCC had equivalent performance as NFL and GCL. The specific spatial locations within the posterior pole that exhibit best performance vary depending on which layer is being assessed. Recognizing this dependency highlights the importance of considering multiple layers independently, as they offer complementary information for effective and comprehensive diagnosis.

Automatic Acne Severity Grading with a Small and Imbalanced Data Set of Low-Resolution Images.

Developing automatic acne vulgaris grading systems based on machine learning is an expensive endeavor in terms of data acquisition. A machine learning practitioner will need to gather high-resolution pictures from a considerable number of different patients, with a well-balanced distribution between acne severity grades and potentially very tedious labeling. We developed a deep learning model to grade acne severity with respect to the Investigator's Global Assessment (IGA) scale that can be trained on low-resolution images, with pictures from a small number of different patients, a strongly imbalanced severity grade distribution and minimal labeling. A total of 1374 triplets of images (frontal and lateral views) from 391 different patients suffering from acne labeled with the IGA severity grade by an expert dermatologist were used to train and validate a deep learning model that predicts the IGA severity grade. On the test set we obtained 66.67% accuracy with an equivalent performance for all grades despite the highly imbalanced severity grade distribution of our database. Importantly, we obtained performance on par with more tedious methods in terms of data acquisition which have the same simple labeling as ours but require either a more balanced severity grade distribution or large numbers of high-resolution images. Our deep learning model demonstrated promising accuracy despite the limited data set on which it was trained, indicating its potential for further development both as an assistance tool for medical practitioners and as a way to provide patients with an immediately available and standardized acne grading tool. chinadrugtrials.org.cn identifier CTR20211314.

Validation of an Artificial Intelligence-Based Prediction Model Using 5 External PET/CT Datasets of Diffuse Large B-Cell Lymphoma.

The aim of this study was to validate a previously developed deep learning model in 5 independent clinical trials. The predictive performance of this model was compared with the international prognostic index (IPI) and 2 models incorporating radiomic PET/CT features (clinical PET and PET models). Methods: In total, 1,132 diffuse large B-cell lymphoma patients were included: 296 for training and 836 for external validation. The primary outcome was 2-y time to progression. The deep learning model was trained on maximum-intensity projections from PET/CT scans. The clinical PET model included metabolic tumor volume, maximum distance from the bulkiest lesion to another lesion, SUVpeak, age, and performance status. The PET model included metabolic tumor volume, maximum distance from the bulkiest lesion to another lesion, and SUVpeak Model performance was assessed using the area under the curve (AUC) and Kaplan-Meier curves. Results: The IPI yielded an AUC of 0.60 on all external data. The deep learning model yielded a significantly higher AUC of 0.66 (P < 0.01). For each individual clinical trial, the model was consistently better than IPI. Radiomic model AUCs remained higher for all clinical trials. The deep learning and clinical PET models showed equivalent performance (AUC, 0.69; P > 0.05). The PET model yielded the highest AUC of all models (AUC, 0.71; P < 0.05). Conclusion: The deep learning model predicted outcome in all trials with a higher performance than IPI and better survival curve separation. This model can predict treatment outcome in diffuse large B-cell lymphoma without tumor delineation but at the cost of a lower prognostic performance than with radiomics.