Search

Breast cancer has high incidences and mortality rates in women worldwide. Malignancy could be detected manually by experienced pathologists based on Hematoxylin and Eosin (H&E) stained images. However, it is time‐consuming and experience‐dependent, making early diagnosis a big challenge. In this paper, a methodology for breast cancer classification based on histopathological images with deep learning was described. A residual learning‐based convolutional neural network named myResNet‐34 was designed for malignancy‐and‐benign classification. In addition, an algorithm automatically generating the target image for stain normalization was proposed, which eliminated the bias caused by manual selection of the reference image. Elastic distortion was introduced and combined with affine transformation for data augmentation considering the characteristics of the H&E images. Experiments were conducted on BreakHis dataset with the proposed framework. Promising results were achieved with an average classification accuracy of around 91% on image‐level classification. Results indicated that both our data augmentation and stain normalization effectively improved the classification accuracy by 2‐3%. [ABSTRACT FROM AUTHOR]

The paper assesses how top incomes biases affect the estimation of income inequality in the United States using the Current Population Survey (the March Annual Social and Economic Supplement) and two alternative correction methods – a stochastic approach based on reweighting and a semi‐parametric approach based on replacing observations. Consistently with previous studies, both methods and their joint application show that income inequality in the United States between 1979 and 2014 has been consistently underestimated by several percentage points. The level of underestimation is positively and significantly associated with mean income, non‐response rates and the initial level of inequality. Reweighting is found to address top incomes biases – specifically those related to unit and item non‐response – consistently and more effectively than replacing, possibly because widely used parametric distributions do not represent US top incomes accurately. [ABSTRACT FROM AUTHOR]

Background: Cone beam computed tomography (CBCT) plays an increasingly important role in image‐guided radiation therapy. However, the image quality of CBCT is severely degraded by excessive scatter contamination, especially in the abdominal region, hindering its further applications in radiation therapy. Purpose: To restore low‐quality CBCT images contaminated by scatter signals, a scatter correction algorithm combining the advantages of convolutional neural networks (CNN) and Swin Transformer is proposed. Methods: In this paper a scatter correction model for CBCT image, the Flip Swin Transformer U‐shape network (FSTUNet) model, is proposed. In this model, the advantages of CNN in texture detail and Swin Transformer in global correlation are used to accurately extract shallow and deep features, respectively. Instead of using the original Swin Transformer tandem structure, we build the Flip Swin Transformer Block to achieve a more powerful inter‐window association extraction. The validity and clinical relevance of the method is demonstrated through extensive experiments on a Monte Carlo (MC) simulation dataset and frequency split dataset generated by a validated method, respectively. Result: Experimental results on the MC simulated dataset show that the root mean square error of images corrected by the method is reduced from over 100 HU to about 7 HU. Both the structural similarity index measure (SSIM) and the universal quality index (UQI) are close to 1. Experimental results on the frequency split dataset demonstrate that the method not only corrects shading artifacts but also exhibits a high degree of structural consistency. In addition, comparison experiments show that FSTUNet outperforms UNet, Deep Residual Convolutional Neural Network (DRCNN), DSENet, Pix2pixGAN, and 3DUnet methods in both qualitative and quantitative metrics. Conclusions: Accurately capturing the features at different levels is greatly beneficial for reconstructing high‐quality scatter‐free images. The proposed FSTUNet method is an effective solution to CBCT scatter correction and has the potential to improve the accuracy of CBCT image‐guided radiation therapy. [ABSTRACT FROM AUTHOR]

Background: Scattering photons can seriously contaminate cone‐beam CT (CBCT) image quality with severe artifacts and substantial degradation of CT value accuracy, which is a major concern limiting the widespread application of CBCT in the medical field. The scatter kernel deconvolution (SKD) method commonly used in clinic requires a Monte Carlo (MC) simulation to determine numerous quality‐related kernel parameters, and it cannot realize intelligent scatter kernel parameter optimization, causing limited accuracy of scatter estimation. Purpose: Aiming at improving the scatter estimation accuracy of the SKD algorithm, an intelligent scatter correction framework integrating the SKD with deep reinforcement learning (DRL) scheme is proposed. Methods: Our method firstly builds a scatter kernel model to iteratively convolve with raw projections, and then the deep Q‐network of the DRL scheme is introduced to intelligently interact with the scatter kernel to achieve a projection adaptive parameter optimization. The potential of the proposed framework is demonstrated on CBCT head and pelvis simulation data and experimental CBCT measurement data. Furthermore, we have implemented the U‐net based scatter estimation approach for comparison. Results: The simulation study demonstrates that the mean absolute percentage error (MAPE) of the proposed method is less than 9.72% and the peak signal‐to‐noise ratio (PSNR) is higher than 23.90 dB, while for the conventional SKD algorithm, the minimum MAPE is 17.92% and the maximum PSNR is 19.32 dB. In the measurement study, we adopt a hardware‐based beam stop array algorithm to obtain the scatter‐free projections as a comparison baseline, and our method can achieve superior performance with MAPE < 17.79% and PSNR > 16.34 dB. Conclusions: In this paper, we propose an intelligent scatter correction framework that integrates the physical scatter kernel model with DRL algorithm, which has the potential to improve the accuracy of the clinical scatter correction method to obtain better CBCT imaging quality. [ABSTRACT FROM AUTHOR]

VAR models are popular to forecast macroeconomic time series. However, the model, the parameters, and the error distribution are rarely known without uncertainty, so bootstrap methods are applied to deal with these sources of uncertainties. In this paper, the performance of the popular forecast Bonferroni cubes based on the Gaussian method and variants of the bootstrap procedure that incorporate error distribution, parameter uncertainty, bias correction, and lag order uncertainty are compared. Monte Carlo simulations suggest that the best performance of bootstrap cubes are obtained when the parameter uncertainty is considered, being the bias and model uncertainties important for long‐run forecast regions in persistent VAR models. Similar conclusions are found in an empirical application based on a four variate system containing US monthly percent changes of the industrial production index, the S&P500 stock market index, its dividend yield, and the unemployment rate. [ABSTRACT FROM AUTHOR]

The equation currently reads as: Deviance residual: HT ht and it should be corrected to read as follows: Deviance residual: HT ht Please note that this is an error in the manuscript and is written correctly in the R script used to conduct this analysis; therefore, it does not impact the results. In the paper by Powell-Romero et al. (2023), there was an error in the equation for deviance residuals. Correction to: Improving the predictability and interpretability of co-occurrence modelling through feature-based joint species distribution ensembles. [Extracted from the article]

The variability of sea surface temperatures (SSTs) is crucial in climate dynamics, influencing marine ecosystems and human activities. This study leverages graph neural networks (GNNs), specifically a GraphSAGE model, to forecast SSTs and their anomalies (SSTAs), focusing on the global scale structure of climatological data. We introduce an improved graph construction technique for SST teleconnection representation. Our results highlight the GraphSAGE model's capability in 1‐month‐ahead global SST and SSTA forecasting, with SST predictions spanning up to 2 years with a recursive approach. Notably, regions with persistent currents exhibited enhanced SSTA predictability, contrasting with equatorial and Antarctic areas. Our GNN outperformed both the persistence model and traditional methods. Additionally, we offer an SST and SSTA graph data set based on ERA5 and a graph generation tool. This GNN case study has shown how the GraphSAGE can be used in SST and SSTA forecasting, and will provide a foundation for further refinements such as adjusting graph construction, optimizing imbalanced regression techniques for extreme SSTAs, and integrating GNNs with other temporal pattern learning methods to improve long‐term predictions. Plain Language Summary: Have you ever wondered how scientists predict climate changes like warming oceans or unusual weather patterns? Our team is improving these predictions by focusing on sea surface temperatures (SSTs) and SST anomalies (SSTAs)—changes in ocean temperatures that greatly impact our environment and society. Traditionally, scientists use grid‐based machine learning models for these forecasts. We have taken a new approach using graph neural networks (GNNs), which are excellent at learning spatial patterns. This method does not just predict temperatures; it understands how different ocean areas are interconnected. Our technique successfully forecasts SSTs up to 2 years ahead and SSTAs 1 month in advance, particularly in regions with strong ocean currents. This accuracy is crucial for anticipating and preparing for climate impacts on marine and human activities. We have developed tools and data sets for researchers, which will foster advancements in graph‐based climate forecasting. Our findings demonstrate that using appropriate graph re‐sampling and GNNs can help understand the complex climate system. In summary, we are pioneering new ways to predict ocean temperature changes through machine learning, contributing to our understanding of and adaptation to climate change. This work benefits not just scientists, but anyone interested in climate and ocean studies. Key Points: Explores the power of graph representation for global sea surface temperature (SST) and SST anomaly (SSTA) forecastingAddresses the issues with traditional grid‐based models and provides an SST graph data set and a tool for converting SST grid data to graphsShows the efficacy of the GraphSAGE model for global SST and SSTA forecasting and identifies the regions where SSTAs are more predictable [ABSTRACT FROM AUTHOR]

Purpose: To investigate the impact of MRI patient‐specific geometrical distortion (PSD) on the quality of Gamma Knife stereotactic radiosurgery (GK‐SRS) plans of the vestibular schwannoma (VS) tumors. Methods and materials: Three open access datasets including the MPI‐Leipzig Mind‐Brain‐Body (318 patients), the slow event‐related fMRI designs dataset (62 patients), and the VS dataset (242 patients) were used. We used first two datasets to train a 3D convolution network to predict the distortion map of third dataset that were then used to calculate and correct the PSD. GK‐SRS plans of VS dataset were used to evaluate dose distribution of PSD‐corrected MRI images. GK‐SRS prescription dose of VS cases was 12 Gy. Geometric and dosimetric discrepancies were assessed between the dose distributions and contours before and after the PSD corrections. Geometry indices were center of the contours, Dice coefficient (DC), Hausdorff distance (HD), and dosimetric indices were Dμ${D_\mu }$, Dmax${D_{max}}$, Dmin${D_{min}}$, and D95%${D_{95{\mathrm{\% }}}}$ doses, target coverage (TC), Paddick's conformity index (PCI), Paddick's gradient index (GI), and homogeneity index (HI). Results: Geometric distortions of about 1.2 mm were observed at the air‐tissue interfaces at the air canal and nasal cavity borders. Average center of the targets was significantly distorted along the frequency encoding direction after the PSD‐correction. Average DC and HD metrics were 0.90 and 2.13 mm. Average Dμ${D_\mu }$, D95%,${D_{95{\mathrm{\% ,}}}}$ and Dmin${D_{min}}$ in Gy significantly increased after PSD correction from 16.85 to 17.25, 12.30 to 12.77, and from 8.98 to 9.92. Dmax${D_{max}}$ did not significantly change after the correction. Average TC and PCI significantly increased from 0.97 to 0.98, and 0.94 to 0.96. Average GI decreased significantly from 2.24 to 2.15 after PSD correction. However, HI did not significantly change after the correction. Conclusion: The proposed method could predict and correct the PSD that indicates the importance of PSD correction before GK‐SRS plans of the VS patients. [ABSTRACT FROM AUTHOR]

A more accurate representation of ice‐phase processes in numerical models necessitates an enhanced understanding of ice‐particle microphysics and their respective formation conditions. Prior in situ measurements have noted distinctive ice‐particle shape characteristics associated with different cloud systems and geographical locations. The recent advancement in ice‐particle backscattering theories enables a more comprehensive exploration of the geographical distribution and seasonal dependence of ice‐particle shape categories than ever before. This exploration is being undertaken for the first time using space‐borne lidar measurements. Distinct geographical preferences were observed for five different ice‐particle categories. Bullets/rosettes were the most common, followed by Voronois, which were especially prevalent in high‐level tropical clouds, and 2D columns, which were commonly found in mid‐ and low‐level clouds. Droxtals were primarily observed in high‐level subtropical regions. The global distribution of ice‐particle types provides valuable insights into the physical processes related to ice cloud particle shape formation, cloud duration, and radiative impacts. Plain Language Summary: An enhanced understanding of ice‐particle microphysics associated with different cloud systems and geographical locations is expected to improve the representation of ice clouds in numerical models for better future climate predictions. For such purpose, space‐borne lidar observations have been intensively used to characterize the global distribution of cloud phases, as well as ice particle types. Latest theoretical studies have indicated the possibility of further decomposing these ice particle types into more specific ice‐particle category information, but had not been applied to actual global observation data. In this work, the geographical distribution of five ice‐particle categories was derived based on the theoretical studies using spaceborne lidar data for the first time. It was found that different ice‐particle categories had a unique geographical preference for occurrence, along with seasonal and height‐dependent characteristics. The global distribution of ice‐particle categories obtained in the present study is expected to be useful for understanding the physical processes related to ice‐particle shape formation and the ice‐particle terminal velocity characteristics relevant to cloud duration. Key Points: For the first time, spaceborne lidar was used to study meticulously the height‐resolved global distribution of five ice‐particle categoriesEach cloud‐particle type displayed a unique geographical preference for occurrence, along with seasonal and height‐dependent characteristicsThe global data on ice‐particle types and particle size from radar‐lidar synergies promises to be valuable for future cloud modeling studies [ABSTRACT FROM AUTHOR]

The species–area relationship (SAR) has over a 150‐year‐long history in ecology, but how its shape and origins vary across scales and organisms remains incompletely understood. This is the first subcontinental freshwater study to examine both these properties of the SAR in a spatially explicit way across major organismal groups (diatoms, insects, and fish) that differ in body size and dispersal capacity. First, to describe the SAR shape, we evaluated the fit of three commonly used models, logarithmic, power, and Michaelis–Menten. Second, we proposed a hierarchical framework to explain the variability in the SAR shape, captured by the parameters of the SAR model. According to this framework, scale and species group were the top predictors of the SAR shape, climatic factors (heterogeneity and median conditions) represented the second predictor level, and metacommunity properties (intraspecific spatial aggregation, γ‐diversity, and species abundance distribution) the third predictor level. We calculated the SAR as a sample‐based rarefaction curve using 60 streams within landscape windows (scales) in the United States, ranging from 160,000 to 6,760,000 km2. First, we found that all models provided good fits (R2 ≥ 0.93), but the frequency of the best‐fitting model was strongly dependent on organism, scale, and metacommunity properties. The Michaelis–Menten model was most common in fish, at the largest scales, and at the highest levels of intraspecific spatial aggregation. The power model was most frequent in diatoms and insects, at smaller scales, and in metacommunities with the lowest evenness. The logarithmic model fit best exclusively at the smallest scales and in species‐poor metacommunities, primarily fish. Second, we tested our framework with the parameters of the most broadly used SAR model, the log–log form of the power model, using a structural equation model. This model supported our framework and revealed that the SAR slope was best predicted by scale‐ and organism‐dependent metacommunity properties, particularly spatial aggregation, whereas the intercept responded most strongly to species group and γ‐diversity. Future research should investigate from the perspective of our framework how shifts in metacommunity properties due to climate change may alter the SAR. [ABSTRACT FROM AUTHOR]

Rear‐edge populations at the warm margin of species distribution are small, isolated and face environmental conditions at the limit of species bioclimatic envelope. Intraspecific phenotypic variation contributing to the persistence of peripheral populations is expected to become increasingly important under future climate conditions to avoid local extirpation where range shifts lag behind climate change velocity.We investigated the putative role of intraspecific phenotypic variation for the maintenance of rear‐edge populations of fire‐prone jack pine (Pinus banksiana), an obligate pyriscent boreal species. We assessed whether variation in cone serotiny is associated with the population trajectory of marginal stands located south of the boreal biome, in the temperate forest where natural wildfires are infrequent and unpredictable. To this end, we estimated stand‐scale serotiny, minimal age and tree size structure in 26 jack pine stands from the rear edge (n = 17 sites) and the core (n = 9 sites) of the species' range in eastern Canada.On average, rear‐edge jack pine populations are less serotinous albeit more variably compared with range‐core populations where serotiny is more uniformly high. Rear‐edge stands are generally older and display reverse J‐shape tree size structure indicative of a multi‐aged demographic equilibrium, whereas range‐core stands are younger and show a unimodal stand structure depicting a single ageing cohort generally lacking interfire recruitment. Eco‐evolutionary dynamics shifts from a dependency on wildfires in range‐core populations to stands that can regenerate and persist without recurrent fires at the rear edge, where stand‐scale serotiny reaches values below 85%.Synthesis. Unlike range‐core populations, rear‐edge jack pine populations can locally rely on a dual life‐history strategy to ensure both steady recruitment during fire‐free intervals and successful postfire regeneration. This capacity to cope with infrequent and unpredictable fire regime should increase the resilience and resistance of jack pine populations as global changes alter fire dynamics of the boreal forest. More generally, unique intraspecific phenotypic variation in rear‐edge populations contributes to long‐term species persistence in marginal environmental conditions that might scale up with global changes. The conservation of rear‐edge populations and their genetic legacy appears crucial for the resilience of species. [ABSTRACT FROM AUTHOR]

Improving precipitation accuracy over a watershed is one of the highest priorities in water resources studies and management. Several global precipitation datasets are available for estimating precipitation over any region in the world. However, local or regional application of these datasets should account for and correct potential errors in the original products. This article presents a novel spatiotemporal calibration framework to improve the accuracy (bias and correlation) of global precipitation datasets in regional applications. The proposed methodology consists of two steps. First, gridded global precipitation datasets are regressed pointwise against rain gauge data. This yields downscaled and bias‐corrected precipitation values at the point scale. Second, the resulting point‐scale regression parameters are used to build a geostatistical model that predicts the regression parameters across the region of interest, allowing for bias‐correcting the precipitation datasets at the regional scale. The framework is applied to the Urmia Lake Watershed in northwestern Iran. Eight global high‐resolution monthly precipitation datasets (CHIRPS, ERA‐5, IMERG, PERSIANN, PERSIANN‐CCS, PERSIANN‐CDR, TRMM and Terra) are evaluated and three downscaling approaches including linear, Q‐Q and Linear Scaling (LS) regression methods are used to calibrate the precipitation datasets based on a regional network of rain gauge observations. Ordinary kriging is subsequently used to predict the regression parameters at ungauged locations. Out of all combinations (i.e., eight datasets and three methods), downscaled IMERG using linear and Q‐Q regression methods showed the best performance in estimating the spatiotemporal variations of monthly precipitation across the watershed of interest. The original IMERG dataset overestimated the monthly precipitation by approximately 20% compared to the precipitation from rain gauges. After applying the proposed methodology in this article, the IMERG bias was reduced by 93%, with an additional 26% decrease in the RMSE. [ABSTRACT FROM AUTHOR]

In this article, we revisit the problem of how to construct better nonparametric confidence intervals for the conditional quantile function from an optimization perspective. We apply the fully data‐driven bias correction procedure based on local polynomial smoothing to estimate the conditional quantile. To account for the effect of the estimated bias, we apply an asymptotic framework that the ratio of the bandwidth to the pilot bandwidth tends to some positive constant rather than zero as the sample size grows. We derive an alternative asymptotic normality of the proposed bias‐corrected quantile estimator as well as a new asymptotic variance formula. Based on theoretical results, two new pointwise confidence intervals are constructed through resampling strategies. Extensive simulation studies show that our proposed confidence intervals enjoy better performance than other competitors in terms of coverage probabilities and interval lengths and are not sensitive to the choice of bandwidth. Finally, our proposed procedure is further illustrated through United States' natality birth data in 2017. [ABSTRACT FROM AUTHOR]

The fidelity of climate projections is often undermined by biases in climate models due to their simplification or misrepresentation of unresolved climate processes. While various bias correction methods have been developed to post‐process model outputs to match observations, existing approaches usually focus on limited, low‐order statistics, or break either the spatiotemporal consistency of the target variable, or its dependency upon model resolved dynamics. We develop a Regularized Adversarial Domain Adaptation (RADA) methodology to overcome these deficiencies, and enhance efficient identification and correction of climate model biases. Instead of pre‐assuming the spatiotemporal characteristics of model biases, we apply discriminative neural networks to distinguish historical climate simulation samples and observation samples. The evidences based on which the discriminative neural networks make distinctions are applied to train the domain adaptation neural networks to bias correct climate simulations. We regularize the domain adaptation neural networks using cycle‐consistent statistical and dynamical constraints. An application to daily precipitation projection over the contiguous United States shows that our methodology can correct all the considered moments of daily precipitation at approximately 1° resolution, ensures spatiotemporal consistency and inter‐field correlations, and can discriminate between different dynamical conditions. Our methodology offers a powerful tool for disentangling model parameterization biases from their interactions with the chaotic evolution of climate dynamics, opening a novel avenue toward big‐data enhanced climate predictions. Plain Language Summary: Accurate climate prediction is crucial for understanding climate change and implementing effective climate adaptation strategies. However, climate models that are used to generate climate predictions have multifaceted biases that often need to be corrected before predictions can be considered usable. We develop a data‐driven methodology that detects and corrects climate model biases using a game theory inspired machine learning technique. By applying physical and statistical constraints, our predictions are not only more accurate, but also more trustworthy as judged by our physical understandings. Key Points: Identify and correct climate projection biases using unpaired climate simulation and observation dataRegularize the data‐driven bias corrector using statistical and dynamical constraintsSignificant improvement of daily precipitation estimation regarding a broad range of spatiotemporal statistics [ABSTRACT FROM AUTHOR]