Search

Accurate detection of small objects plays an important role in the application of Autonomous aerial vehicles (AAV). However, current works mainly extract comprehensive features from unimodal images, which can obtain very limited distinguishable features for objects, especially those with small sizes. To address this issue, we propose a dynamic cascade cross-modal coassisted network, which integrates multimodal images fusion and fine-grained feature learning to generate powerful object semantic representations. Specifically, we design a multimodal high-order interaction module to achieve collaborative interaction of spatial details and channel dependencies between modalities, thereby enhancing object discrimination. To preserve multimodal fine-grained details, we devise a scale-adaptive dynamic feature prompt module, which dynamically motivates the backbone network to capture feature degradation clues. Meanwhile, to maintain the spatial correlation of multimodal cross-scale features and improve the quality of feature fusion, we derive a global collaborative enhancement module into the feature pyramid network for enhancing the detection accuracy across multiple scales. Extensive experimental results on multimodal datasets have shown that our method achieves favorable performance, surpassing other state-of-the-art methods.

Abstract Background Autophagy is a crucial mechanism for maintaining cellular homeostasis and responding to environmental stress, and it is closely linked to tumor drug resistance. Through multi-omics analysis, this study explores the expression patterns, functions, and potential role of the autophagy-related gene Angiotensinogen (AGT) in colorectal cancer (CRC), particularly in relation to chemotherapy resistance. Methods This study first compared AGT expression between CRC and normal tissues using the GTEx and TCGA databases. Differences in expression were assessed using Wilcoxon Rank Sum Tests, and the prognostic impact of AGT was evaluated through univariate Cox survival analysis and meta-analysis. Functional enrichment was performed using the limma and fgsea packages. Drug sensitivity analysis was conducted based on the CTRP database, while immune infiltration was assessed using the CIBERSORT and ESTIMATE methods. Spatial transcriptomic characteristics were explored through 10x Visium technology and deconvolution analysis to investigate the correlation between AGT expression levels and tumor cell content.scRNA-seq data from CRC tissues were sourced from Tumor Immune Single Cell Hub (TISCH).Functional annotation was performed with Single-sample gene set enrichment analysis (SSGSEA), and pseudotime analysis using Monocle 2 mapped their developmental trajectories. The potential of AGT inhibitors in the treatment of CRC was analyzed using drug-target Mendelian randomization.Finally, Phenome-Wide Association Study (PheWAS) was conducted to evaluate genetic associations and potential side effects of AGT inhibitors. Results AGT expression was significantly higher in CRC tissues compared to normal tissues and was associated with shorter recurrence-free survival (RFS). Autophagy signaling pathways were markedly enriched in the high AGT expression group. AGT expression was positively correlated with resistance to chemotherapeutic agents such as gemcitabine, cisplatin, paclitaxel, and 5-fluorouracil. Spatial transcriptomic analysis revealed that AGT was predominantly expressed in malignant tumor regions. Single-cell analysis identified 21 distinct cell subpopulations across 13 major types. AGT expression was significantly higher in tumor samples, especially in the fibroblast C6 subpopulation. Tumor-related pathways were enriched in C1, C5, C6, and C8 subpopulations. Pseudotime analysis revealed that these subpopulations, particularly C6, were in terminal developmental stages.Drug-target Mendelian randomization analysis indicated a negative causal relationship between AGT inhibitors and the risk of both heart failure(ORdrug = 0.950, 95% CI, 0.912–0.990; P = 0.014) and CRC(ORdrug = 0.874, 95% CI: 0.792–0.964; P = 0.007).PheWAS analysis showed no genetic associations between AGT inhibitors and other traits, indicating its specificity and low risk of side effects. Conclusion Elevated AGT expression in CRC is associated with resistance to chemotherapy, and its inhibition may offer a therapeutic avenue for the treatment of colorectal cancer.

Abstract Accurate diagnosis of white blood cells from cytopathological images is a crucial step in evaluating leukaemia. In recent years, image classification methods based on fully convolutional networks have drawn extensive attention and achieved competitive performance in medical image classification. In this paper, we propose a white blood cell classification network called ResNeXt-CC for cytopathological images. First, we transform cytopathological images from the RGB color space to the HSV color space so as to precisely extract the texture features, color changes and other details of white blood cells. Second, since cell classification primarily relies on distinguishing local characteristics, we design a cross-layer deep-feature fusion module to enhance our ability to extract discriminative information. Third, the efficient attention mechanism based on the ECANet module is used to promote the feature extraction capability of cell details. Finally, we combine the modified softmax loss function and the central loss function to train the network, thereby effectively addressing the problem of class imbalance and improving the network performance. The experimental results on the C-NMC 2019 dataset show that our proposed method manifests obvious advantages over the existing classification methods, including ResNet-50, Inception-V3, Densenet121, VGG16, Cross ViT, Token-to-Token ViT, Deep ViT, and simple ViT about 5.5–20.43% accuracy, 3.6–23.56% F1-score, 3.5–25.71% AUROC and 8.1–36.98% specificity, respectively.

Study region: Two boundary-scale watersheds (Yangjiagou and Dongzhuanggou) and three typical vegetation plots [a single plantation, a mixed plantation, and a natural grassland] in the Loess Plateau. Study focus: The performance of different Generalized Complementary Relationships (four versions of GCR) in simulating actual evapotranspiration (ETa) at boundary (∼1 km2) and plot scales (much less than 1 km2) remains unclear, particularly under heterogeneous vegetation cover. Furthermore, the feasibility of using GCR directly to obtain accurate net ETa (ETan, the sum of soil evaporation and plant transpiration) through parameter calibration is investigated. The influence of timescale and vegetation diversity is also discussed. New hydrological insights for the region: Dynamic Scaling of the GCR could achieve high accuracy at boundary-scale watersheds (NSE=0.78) and different vegetation plots (NSE=0.79), indicating the universality and stability of the GCR, in line with the conclusions of previous large-scale studies. Reducing the timescale will decrease the GCR's simulation performance for both ETa and ETan. Interestingly, as vegetation heterogeneity increased, GCR's performance improved when simulating ETa but decreased when simulating ETan. Furthermore, using GCR to obtain ETan directly may generate bias near extreme points at a weekly timescale but becomes feasible at longer timescales. This study expanded recognition of GCR's performance under varying spatiotemporal scales and provided evidence for the possibility of simulating evapotranspiration components by the method.

BackgroundHigh interleukin-6 levels correlate with diseases like cancer, autoimmune disorders, and infections. IL-6 receptor inhibitors (IL-6Ri), used for rheumatoid arthritis and COVID-19, may have wider uses. We apply drug-target Mendelian Randomization (MR) to study IL-6Ri’s effects.MethodTo simulate the effects of genetically blocking the IL-6R, we selected single nucleotide polymorphisms (SNPs) within or near the IL6R gene that show significant genome-wide associations with C-reactive protein. Using rheumatoid arthritis and COVID-19 as positive controls, our primary research outcomes included the risk of asthma, asthmatic pneumonia, cor pulmonale, non-small cell lung cancer, small cell lung cancer, Parkinson’s disease, Alzheimer’s disease, ulcerative colitis, Crohn’s disease, systemic lupus erythematosus, type 1 diabetes, and type 2 diabetes. The Inverse Variance Weighted (IVW) method served as our principal analytical approach, with the hypotheses of MR being evaluated through sensitivity and colocalization analyses. Additionally, we conducted Bayesian Mendelian Randomization analyses to minimize confounding and reverse causation biases to the greatest extent possible.ResultsIL-6 inhibitors significantly reduced the risk of idiopathic pulmonary fibrosis (OR= 0.278, 95% [CI], 0.138–0.558; P 0.05) and were excluded from further analysis. Similarly, Small cell lung cancer were excluded due to inconsistent results. Notably, the colocalization evidence for asthmatic pneumonia (coloc.abf-PPH4 = 0.811) robustly supports its association with CRP. The colocalization evidence for Parkinson’s disease (coloc.abf-PPH4 = 0.725) moderately supports its association with CRP.ConclusionIL-6Ri may represent a promising therapeutic avenue for idiopathic pulmonary fibrosis, Parkinson’s disease, and Type 2 diabetes.

Human activity recognition by radar sensors plays an important role in healthcare and smart homes. However, labeling a large number of radar datasets is difficult and time-consuming, and it is difficult for models trained on insufficient labeled data to obtain exact classification results. In this paper, we propose a multiscale residual weighted classification network with large-scale, medium-scale, and small-scale residual networks. Firstly, an MRW image encoder is used to extract salient feature representations from all time-Doppler images through contrastive learning. This can extract the representative vector of each image and also obtain the pre-training parameters of the MRW image encoder. During the pre-training process, large-scale residual networks, medium-scale residual networks, and small-scale residual networks are used to extract global information, texture information, and semantic information, respectively. Moreover, the time–channel weighting mechanism can allocate weights to important time and channel dimensions to achieve more effective extraction of feature information. The model parameters obtained from pre-training are frozen, and the classifier is added to the backend. Finally, the classifier is fine-tuned using a small amount of labeled data. In addition, we constructed a new dataset with eight dangerous activities. The proposed MRW-CN model was trained on this dataset and achieved a classification accuracy of 96.9%. We demonstrated that our method achieves state-of-the-art performance. The ablation analysis also demonstrated the role of multi-scale convolutional kernels and time–channel weighting mechanisms in classification.

The KCNQ1 gene encodes a voltage-gated potassium channel, which plays an important role in the repolarization of myocardial action potentials. Mutations in this gene often result in type 1 long QT syndrome (LQT1). Here, we generated a KCNQ1 (c.1032 + 2 T > C) mutant human embryonic stem cell line (WAe009-A-1D) based on the transient expression adenine base editing system that converts base A to G. The WAe009-A-1D cell maintains the morphology, pluripotency, and normal karyotype of the stem cells and is capable of differentiating into all three germ layers in vivo.

Bilateral upper urinary tract stones are significantly related to renal function damage. However, few studies characterized the risk factors of bilateral upper urinary tract stones. We retrospectively enrolled 3905 patients with urinary tract stones from March 2019 to March 2022 at the Second Hospital of Tianjin Medical University. Patients were divided into two groups according to the location of the stones, and the related data were evaluated. In this study, 2485 unilateral and 1420 bilateral stone patients were included. Multivariate logistic regression analysis showed that BMI, gout, hyperparathyroidism, uric acid stone, urine PH, 24-h urinary calcium, blood uric acid, and metabolic syndrome (Mets) were independent risk factors for bilateral stone formation(P

Remote sensing image change detection (RSCD) based on deep learning technology has made remarkable achievements. Meanwhile, the enhancement of network architectures and advancements in optimization algorithms have pushed RSCD performance to a higher stage. However, the existing RSCD methods mainly focus on extracting differential information between pixel pairs from bitemporal images, while neglecting the importance of using complementary multiscale features to uncover hierarchical semantic change information. To address these issues, we propose a multiscale feature interaction hybrid network (MFIHNet), which aims to enhance feature discriminability by multiscale features interaction and fusion to enhance model RSCD performance. Specifically, we first design a cascaded convolutional neural networks (CNN)-Transformer feature extraction network to capture hierarchical features at different scales. This strategy enables the network to preserve detailed information in shallow layers while grasping more contextual information in high layers. Subsequently, based on the differences between hierarchical features, we design a novel edge enhancement module (EM) to adaptively focus on key areas under the guidance of edge information to make the changed information clearer. Furthermore, to ensure complementary advantages among different feature layers, we devise a novel cross-scale feature interaction module, which introduces a region-specific atrous convolution into the multiscale attention mechanism for improving feature coassistance capacity. In this way, the MFIHNet not only effectively obtains different types of fine-grained information but also reduces the loss incurred during the feature fusion process, thereby improving the performance of remote sensing RSCD tasks. Extensive experimental results on the challenging CDD and GZ-CD datasets, with mean F1 scores reaching 98.1% and 87.4%, respectively, demonstrate that the proposed method achieves competitive performance. Our source codes are available at https://github.com/qliu520/MFIHNet.

Abstract To further evaluate the causal relationships between inflammatory cytokines and migraine, we conducted a bidirectional, two-sample Mendelian randomization (MR) analysis using genetic data from publicly available genome-wide association studies (GWAS). We used several MR methods, including random-effect inverse-variance weighting (IVW), weighted median, MR-Egger, to test the causal relationships. Sensitivity analyses were also conducted to evaluate the robustness of the results. The results showed that hepatocyte growth factor (HGF) was positively associated with the risk of migraine (odds ratio [OR], 1.004; 95% confidence interval [CI], 1.001–1.008; P = 0.022). In addition, Interleukin-2 (IL-2) was considered a downstream consequence of migraine (OR, 0.012; 95% CI, 0.000–0.0929; P = 0.046). These findings suggest that HGF may be a factor associated with the etiology of migraine, while IL-2 is more likely to be involved in the downstream development of migraine.

Background: In the realm of tumor-targeted therapeutics, Polo-like kinases (PLKs) are a significant group of protein kinases that were found recently as being related to tumors. However, the significance of PLKs in pan-cancer remains systematically studied. Methods and materials: We integrated multi-omics data to comprehensively investigate the expression patterns of the PLK family across various cancer types. Subsequently, study examined the associations between tumor mutation burden (TMB), microsatellite instability (MSI), immune subtype classification, immune infiltration, tumor microenvironment scores, immune checkpoint gene expression, and the PLKs expression profiles within various tumor types. Furthermore, using our mRNA sequencing data (TRUCE01) and four bladder cancer (BLCA) cohorts (GSE111636, GSE176307, and IMvigor210), We examined the correlation between the expression level of PLK and immunotherapy effectiveness. Next, Gene set enrichment analysis (GSEA) was evaluated to find that potentially enriched PLK signaling pathways. Utilizing TIMER 2.0, we conducted an immune infiltration analysis underlying transcriptome expression, copy number variations (CNV), or somatic mutations of PLKs in BLCA. Finally, mRNA expression validation of PLK1/3/4 by real-time PCR within 10 paired BLCA tissues, protein expression verification through the Human Protein Atlas (HPA), and PLK4 in vitro cytological studies have been employed in BLCA. Results: The expression of most of the PLK family members exhibits variation between cancerous tissues and adjacent normal tissues across various cancer species. Furthermore, the expression of PLKs demonstrates a significant association with immunotyping, infiltration of immune cell, tumor mutational burden (TMB), microsatellite instability (MSI), immunological checkpoint gene activity and therapeutic effectiveness in pan-tumor tissues. Additional investigation into the correlation between the PLK family and BLCA has revealed that the expression of the PLK genes holds substantial significance in the biological processes of BLCA. Furthermore, a notable association has been observed between the copy number variation, variant status, and the degree of certain immunological cell infiltration. Of note, the expression validation and in vitro phenotypic experiments have demonstrated that PLK4 has a significant function in promoting the BLCA cell proliferation, migration, and invasion. Conclusion: Collectively, based on various databases, our results highlight the involvement of PLK gene family in the formation of different types of tumors and identify PLK-related genes that may be used for therapy.

Quantifying the evapotranspiration (ET) water consumption characteristics of typical vegetation [Robinia pseudoacacia (single vegetation), mixed plantations of Robinia pseudoacacia and Platycladus orientalis (double-mixed vegetation), and natural grassland (multiple-mixed vegetation)] in the Chinese Loess Plateau has both theoretical and practical significance. However, the proportionality and inverse hypotheses of ET, namely, the FAO-56 method and generalized complementary relationship (GCR), have limitations when used individually to estimate ET, evaporation (E), and transpiration (T). This study explores the feasibility of scaling water stress coefficient (Ks) using calibration-free or parameterized GCR for different vegetation types and quantities. The single crop coefficient (Kc) curves and basal crop coefficient adjustment (Kcb-a/Kcb-p) curves of typical vegetation were defined and validated to partition ET without soil moisture usage. The results revealed that both curves exhibited polynomial functional response relationships to the week ordinal (WO) of the growing season within a complete hydrological cycle (with 0.60 < R2 < 0.78 at the 0.01 significance level). By expressing the soil evaporation coefficient (Ke) as a function of Ks, Kc, and Kcb, the calibration-free combined method (combining the GCR and the FAO-56 method – GCR/FAO-56) demonstrated superior simulation performance for E, T, and ET across different vegetation types with an average Nash–Sutcliff Efficiency of 0.71 during the validation period compared to that of the parameterized method, with the highest sensitivity in E simulation. The simulated E, T, and ET values closely matched the actual values and maintained the order of Robinia pseudoacacia > mixed plantation > natural grassland during the growing season. The average biases for E, T, and ET during the validation period were 18.5, 7.0, and 25.5 mm, respectively, in Robinia pseudoacacia; 10.8, 11.1, and 19.7 mm in mixed plantation; and 2.7, 7.0, and 19.6 mm in natural grassland. The overall T/ET ratios for both plantations exceeded 51.07%, with mixed plantation being higher, surpassing 70% in dry years and exhibiting minimal variation in T (44.4 mm) across hydrological years. However, T/ET ratios in natural grassland were < 0.5, indicating that E was the primary form of water consumption. Average relative biases of the simulated and actual T/ET values were only 1.76%, 2.53%, and 1.87% for Robinia pseudoacacia, mixed plantation, and natural grassland, respectively. Furthermore, the accuracy of the calibration-free combined method increased with the number of vegetation species during the validation period. This study extended and discussed the feasibility in typical vegetation of combining the positive and negative correlation principles of evapotranspiration and further provided a possibility for how to partition ET with a fewer dataset [Rn, u, Tavg, RH or Td, P, LAI] without soil moisture, and also provided scientific suggestions for future vegetation restoration as well as water resources management.

The Hetao Irrigation District (HID) is a typical seasonal frozen soil region in Inner Mongolia. Soil salinization is directly linked to the freeze-thaw cycle of soil water-heat-salt transport. Autumn irrigation is critical for water storage and salt washing, which can effectively prevent and control soil salination. However, less attention has been given to the spatial and temporal variations of soil moisture, heat, and salt throughout the freeze-thaw phase under autumn irrigation at the HID regional scale. Therefore, the distributed SHAW model was constructed, calibrated, and validated, and then used to simulate soil water-heat-salt transport during the freeze-thaw periods under autumn irrigation in the HID from 2000 to 2017. Subsequently, the suitability of planting spring wheat, spring maize and sunflowers was quantitatively determined based on the soil salt content at the time of spring sowing. The results indicated that the distributed SHAW model was effective in simulating the soil water-heat-salt dynamics in the HID. Within the year of freeze-thaw period, the 0–40 cm soil temperature and water content exhibited a similar “plunge-stabilization-recovery” trend, the 0–40 cm soil salinity during the ablation period increased by 5.68% compared to the initial freeze period. During the freeze-thaw period from 2000 to 2017, soil water content showed an increasing trend, soil temperature variability was stable, and soil salinity was desalinated. The spatial distribution of soil water content during the freeze-thaw period was higher in the east than in the west, ranging from 0.07 to 0.13 cm3/cm3. The soil temperature was relatively obvious, ranging from −0.8 to −2.2 ℃, with high soil temperature mainly concentrated in Dengkou county (DK). The soil salinity varied significantly, with most cultivated land falling into the lightly salted soil category with salinity of 2.1 - 2.3 g/kg, while all soils had less than 1.5 g/kg salinity and were non-salinized soils in Wulateqianqi (WQQ). Under existing autumn irrigation, the area suitable for spring wheat cultivation was mainly concentrated in WQQ, accounting for approximately 28.36%; the areas suitable for planting spring maize were mainly in DK, the northern part of Hangqinhouqi (HH), Wuyuan county (WY), and WQQ, accounting for approximately 76.09%; all the cultivated land was satisfied with sunflower cultivation in the HID. The research results provide valuable references for adjusting and optimizing autumn irrigation and cropping patterns in the HID.

This study used feces from 0-day-old (36 rabbits), 10-day-old (119 rabbits), and 60-day-old (119 rabbits) offspring rabbits and their corresponding female rabbits (36 rabbits) as experimental materials. Using 16s rRNA sequencing, the study analyzed the types and changes of gut microbiota in rabbits at different growth and development stages, as well as the correlation between gut microbiota composition and the weight of 60-day-old rabbits. All experimental rabbits were placed in the same rabbit shed. Juvenile rabbits were fed solid feed at 18 days of age and weaned at 35 days of age. In addition to identifying the dominant bacterial phyla of gut microbiota in rabbits at different age stages, it was found that the abundance of Clostridium tertium and Clostridium paraputrificum in all suckling rabbits (10-day-old) was significantly higher than that in rabbits fed with whole feed (60-day-old) (p < 0.05), while the abundance of Gram-negative bacterium cTPY13 was significantly lower (p < 0.05). In addition, Fast Expected Maximum Microbial Source Tracing (FEAST) analysis showed that the contribution of female rabbits’ gut microbiota to the colonization of offspring rabbits’ gut microbiota was significantly higher than that of unrelated rabbits’ gut microbiota (p < 0.05). The contribution of female rabbits’ gut microbiota to the colonization of gut microbiota in 0-day-old rabbits was significantly higher than that to the colonization of gut microbiota in the 10- and 60-day-old rabbits (p < 0.05). Finally, the correlation between gut microbiota composition and body weight of 60-day-old rabbits was analyzed based on a mixed linear model, and six ASVs significantly affecting body weight were screened. The above results provide important theoretical and practical guidance for maintaining gut health, improving growth and development performance, and feeding formulation in rabbits.

We presented a novel hardware architecture that uses dual Benes networks to accelerate Convolutional Neural Network (CNN) algorithms. This architecture can reduce the need for high-speed buses and maintain a high-speed connection between execution units and memories. Also, this design can work with multiple neural network models by changing the Benes network configuration only due to the reorderability of the non-blocking networks. The proposed architecture can save the time of re-implementing the hardware design. Thus, this architecture can act as a general CNN accelerator. We successfully implemented our CNN accelerator on the Xilinx Virtex UltraScale+ VU19P and evaluated it by running the CNN model LeNet-5.

ObjectiveA string-type mooring system is proposed for the permanent berthing demands of special offshore platforms. So it is necessary to study the hydrodynamic performance of the berthing platform using this mooring system to verify its adaptability to the permanent berthing demands.MethodsBased on the three-dimensional potential flow theory, a numerical model of the platform berthed alongside a terminal wharf is established using the boundary element method, and the frequency domain numerical simulation of the moored platform at different tide levels is carried out. Two mooring systems, i.e., string-type and traditional dock mooring, are used to simulate the hydrodynamic performance of the platform in the time domain under extreme wave conditions, and parameterized analysis is conducted to analyze the influence of different tide levels on the motion response of the platform.ResultsBy providing omni-directional restoring force, the string-type mooring system strengthens the rotational motion limitations of the platform, greatly reduce the most violent rolling motion, transfers the energy of environmental load to translational motion, and further strengthens the restriction of the overall motion of the platform; the motion range of the platform is balanced, stable and controllable; and the effects of the system on the motion limit of the platform are less affected by the change in tide level.ConclusionCompared with the traditional mooring system, the string-type mooring system is more suitable for the permanent berthing of special offshore platforms.

Exponentially increasing water demand and increasingly uncertain hydrological conditions due to climate change (CC) and land use/cover change (LUCC) undoubtedly pose significant challenges to the sustainability of water systems in arid and semi-arid regions worldwide, and even more so in the rain-fed agricultural area of the Loess Plateau (LP) in China, where green water supports all ecological landscapes and most agriculture. Therefore, an accurate assessment of green water resources is crucial for regional development. This paper constructed a comprehensive framework for assessing green water resources based on the MIKE SHE model, defined a new concept of green water components (GWCs) that can be extracted based on existing studies, proposes the potential of green water (GWP), and present a research methodology. To validate this research framework, we applied it to a typical basin in the LP gully area of the Nanxiaohegou (NXHG) basin. The model was calibrated using runoff data from basin outlet stations to extract four GWCs: high-efficiency green water (GWH), low-efficiency green water (GWL), unused green water (GWU), ineffective green water (GWI). We then assessed the GWP of the basin and analyzed its trends, consistency, and dependence on changing environments. The assessment framework is simple and easy to use, providing valuable information for regional revegetation and basin water management, and is highly recommended for replication in other regions.

Identifying the unique change characteristics of evapotranspiration (ET) in the typical vegetation of the semi-arid Loess Plateau and accurately estimating their components, namely evaporation (E) and transpiration (T), poses a challenging work. The typical vegetation consists of Robinia pseudoacacia (RP), Platycladus orientalis (PO), Mixed plantation of RP and PO (MP), Constructed grassland (CG), and Natural grassland (NG). This study constructed a practical framework for ET partition by utilizing the Generalized Complementary Relationship (GCR) and the Modified Priestley-Taylor Equation (MPTE) to estimate actual ET and T, respectively, and E can be obtained as the difference between ET and T. The results indicated that the Dynamic Scaling of the GCR (S2017) performed with the highest accuracy at watershed scales by simultaneously adjusting the non-α parameters and assigning different fixed αe values in different hydrological years. Its simulated values (ETa17) exhibited an extremely significant sigmoid function relationship with the actual ET of different vegetation (ET’aw) at plot scales (R2 > 0.88). The Priestley-Taylor coefficient, αc, used to estimate T of plantations and grasslands, showed significant logarithmic and quadratic polynomial response functions to precipitation (P) (R2 > 0.75), respectively. The combined method displayed high accuracy in estimating ET and its components for plantations and grasslands, with NSE exceeding 0.7 for the whole period. It is noteworthy that this level of performance could be achieved with minimal data requirements [Rn, u, Ta, RH or Td, P, LAI]. The ET’aw and its components increased with the moisture of the environment on the whole but were inconsistent with P on weekly scales. The variations of soil water storage (ΔW) showed negative in single plantations and positive in grasslands, while close to balance in the mixed plantation in one hydrological cycle (CG > NG > MP ≈ 0 > PO > RP). And ΔW of the whole period (2015–2020) showed negative in single plantations and positive in grasslands (CG larger) and mixed plantation. The order of T/ET was MP > RP > PO > CG > NG, with MP and CG maintaining relatively stable T in different hydrological years. These findings suggested that with reasonable control measures, a mixed planting plan of mixed plantation and constructed grassland may be an excellent solution for alleviating the pressure of water resources management and fostering ecological restoration in the future.

IntroductionAccurate white blood cells segmentation from cytopathological images is crucial for evaluating leukemia. However, segmentation is difficult in clinical practice. Given the very large numbers of cytopathological images to be processed, diagnosis becomes cumbersome and time consuming, and diagnostic accuracy is also closely related to experts' experience, fatigue and mood and so on. Besides, fully automatic white blood cells segmentation is challenging for several reasons. There exists cell deformation, blurred cell boundaries, and cell color differences, cells overlapping or adhesion.MethodsThe proposed method improves the feature representation capability of the network while reducing parameters and computational redundancy by utilizing the feature reuse of Ghost module to reconstruct a lightweight backbone network. Additionally, a dual-stream feature fusion network (DFFN) based on the feature pyramid network is designed to enhance detailed information acquisition. Furthermore, a dual-domain attention module (DDAM) is developed to extract global features from both frequency and spatial domains simultaneously, resulting in better cell segmentation performance.ResultsExperimental results on ALL-IDB and BCCD datasets demonstrate that our method outperforms existing instance segmentation networks such as Mask R-CNN, PointRend, MS R-CNN, SOLOv2, and YOLACT with an average precision (AP) of 87.41%, while significantly reducing parameters and computational cost.DiscussionOur method is significantly better than the current state-of-the-art single-stage methods in terms of both the number of parameters and FLOPs, and our method has the best performance among all compared methods. However, the performance of our method is still lower than the two-stage instance segmentation algorithms. in future work, how to design a more lightweight network model while ensuring a good accuracy will become an important problem.

Backgroud: We aimed to explore the prognostic features of ligand and receptor genes associated with disulfidoptosis in hepatocellular carcinoma (HCC) and establish a risk signature utilizing these genes to predict the prognosis of HCC patients. Methods: We used scRNA-seq data from GSE166635 to differentiate malignant cells from normal cells using “copykat”.The study thoroughly examined the disparities in disulfidoptosis scores and the associated gene expressions between malignant and normal cells.We identified key ligand and receptor genes that are specific to HCC cells.Subsequently, Correlation analysis was conducted to ascertain the ligand and receptor genes associated with disulfidoptosis.We performed univariate Cox regression analysis to identify prognostic ligand and receptor genes associated with disulfidoptosis.We employed LASSO to construct a risk signature using prognostic ligand and receptor genes associated with disulfidoptosis.Lastly, we developed a nomogram model that integrates the risk signature with clinicopathological characteristics. Results: Malignant cells displayed a marked increase in disulfidoptosis scores and the expression of associated genes compared to normal cells.We identified 110 receptor and ligand genes significantly associated with disulfidoptosis, and narrowed them down to create a risk signature comprising eight genes.Multivariate analysis confirmed the risk signature as an independent prognostic factor for HCC and validated its predictive value for immunotherapy outcomes.A novel nomogram was developed, incorporating stage information and the risk signature derived from disulfidoptosis-related receptor and ligand genes, demonstrating excellent predictive accuracy and reliability in HCC prognosis prediction. Conclusion: Risk signatures based on disulfidoptosis-associated ligand and receptor genes can effectively predict HCC prognosis and may inform immunotherapy strategies.

Crowd movement analysis (CMA) is a key technology in the field of public safety. This technology provides reference for identifying potential hazards in public places by analyzing crowd aggregation and dispersion behavior. Traditional video processing techniques are susceptible to factors such as environmental lighting and depth of field when analyzing crowd movements, so cannot accurately locate the source of events. Radar, on the other hand, offers all-weather distance and angle measurements, effectively compensating for the shortcomings of video surveillance. This paper proposes a crowd motion analysis method based on radar particle flow (RPF). Firstly, radar particle flow is extracted from adjacent frames of millimeter-wave radar point sets by utilizing the optical flow method. Then, a new concept of micro-source is defined to describe whether any two RPF vectors originated from or reach the same location. Finally, in each local area, the internal micro-sources are counted to form a local diffusion potential, which characterizes the movement state of the crowd. The proposed algorithm is validated in real scenarios. By analyzing and processing radar data on aggregation, dispersion, and normal movements, the algorithm is able to effectively identify these movements with an accuracy rate of no less than 88%.

Straw mulching is a key method for controlling soil and water losses. Mulching costs may be reduced by applying it in strips rather than over entire areas. However, the effect of different straw mulching methods on the effectiveness of reducing soil erosion is unclear. In this study, the effects of straw mulching strip length (covering 1/4, 1/2, 3/4, and 4/4 of the slope length) and coverage rate (0.2, 0.5, and 0.8 kg m−2) on interception, infiltration, runoff, and soil erosion were investigated at the plot scale using rainfall simulation experiments. The further complex correlations between these variables were analyzed using structural equation modeling (SEM). Bare slopes were used as a control group. The rainfall intensity was chosen to be 60 mm h−1. The results showed that (1) the modified Merriam interception model can describe the change in interception with time under straw mulching conditions well (R2 > 0.91, NSE > 0.75). (2) A total of 35.39–78.79% of the rainwater is converted into infiltration on straw-covered slopes, while this proportion is 36.75% on bare slopes. The proportion of rainwater converted to infiltration was greatest (78.79%) when the straw covered 3/4 of the slope length at a coverage rate of 0.5 kg m−2, which was the most conducive to rainwater harvesting on the slope. (3) Straw mulching protects the topsoil from the impact of raindrops and directly affects the sediment yield (direct effect = −0.44). Straw mulching can also indirectly affect sediment yield by increasing interception, reducing runoff, and decreasing the sediment carrying capacity of runoff (indirect effect = −0.83). Compared with bare slopes, straw covering at least 1/2 of the slope length can significantly reduce runoff yield, but straw covering only 1/4 of the slope length can significantly reduce sediment yield. Moreover, once the straw mulch slope length reaches 3/4 and the coverage rate reaches 0.5 kg m−2, further increases in mulch slope length and coverage rate will not significantly reduce the runoff and sediment yields. These results assessed the effectiveness of different straw mulching methods in controlling soil and water losses on sloping farmland.

An exhaust gas recirculation (EGR) valve is used to quickly and dynamically adjust the amount of recirculated exhaust gas, which is critical for improving engine fuel economy and reducing emissions. To address problems relating to the precise positioning of an electromotive (EM) valve under slowly varying plant dynamics and uncertain disturbances, we propose a servo control system design based on linear active disturbance rejection control (LADRC) for the EGR EM valve driven by a limited angle torque motor (LATM). By analyzing the structure of the LATM and the transmission, the dynamic model of the system is derived. In addition, to solve the problems caused by slowly varying plant dynamics and uncertain disturbances, we combine the effects of uncertain model parameters and external disturbances as the total disturbance, which is estimated in real time by an extended state observer (ESO) and then compensated. In addition, accurate angular information is obtained using a non-contact magnetic angle measurement method, and a high-speed digital communication channel is established to help implement a closed-loop position control system with improved responsiveness and accuracy. Simulation and experimental results show that the proposed servo system design can effectively ensure the precision and real-time performance of the EM valve under slowly changing plant dynamics and uncertain disturbances. The proposed servo system design achieves a full-stroke valve control accuracy of better than 0.05 mm and a full-stroke response time of less than 100 ms. The controlled valve also has good robustness under shock-type external disturbances and excellent airflow control capability. The repeatability of the airflow control is generally within 5%, and the standard deviation is less than 0.2 m3/h.

The Protein phosphatase 1 catalytic subunit alpha (PPP1CA) is a alpha subunit of the PP1 complex, which known to be involved in the regulation of a variety of cellular processes, such as cell division, glycogen metabolism, muscle contractility, protein synthesis, and HIV-1 viral transcription. Increased PP1 activity has been observed in the end stage of heart failure. Here, a PPP1CA knockout human embryonic stem cell line, WAe009-A-B, was generated using the CRISPR/Cas9 system to further study the function of PPP1CA. The cell line confirmed with pluripotency, normal karyotype and differentiation potential.

Background: Transient receptor potential cation channel subfamily V (TRPV) play an essential in cancer initiation, progression, and treatment. TRPV expression alteration are shown relate to multiple cancers prognosis and treatment of cancers but are less-studied in pan-cancer. In this study, we characterize the clinical prediction value of TRPV at pan-cancer level. Methods: Several databases were used to examine the transcript expression difference in tumor vs. normal tissue, copy-number variant (CNV) and single nucleotide polymorphisms (SNP) mutation of each TRPV members in pan-cancer, including The Cancer Genome Atlas (TCGA) and cBioPortal. We performed K-M survival curve and univariate Cox regression analyses to identify survival and prognosis value of TRPV. CellMiner were selected to explore drug sensitivity. We also analyzed association between tumor mutation burden (TMB), microsatellite instability (MSI), tumor immune microenvironment and TRPV family genes expression. Moreover, we investigated the relationship between TRPVs expression and effectiveness of immunotherapy in multiple cohorts, including one melanoma (GSE78220), one renal cell carcinoma (GSE67501), and three bladder cancer cohorts (GSE111636, IMvigor210, GSE176307 and our own sequencing dataset (TRUCE-01)), and further analyzed the changes of TRPVs expression before and after treatment (tislelizumab combined with nab-paclitaxel) of bladder cancer. Next, we made a special effort to investigate and study biological functions of TRPV in bladder cancer using gene set enrichment analysis (GSEA), and conducted immune infiltration analysis with TRPVs family genes expression, copy number or somatic mutations of bladder cancer by TIMER 2.0. Finally, real-time PCR and protein expression validation of TRPVs within 10 paired cancer and para-carcinoma tissue samples, were also performed in bladder cancer. Results: Only TRPV2 expression was lower in most cancer types among TRPV family genes. All TRPVs were correlated with survival changes. Amplification was the significant gene alternation in all TRPVs. Next, analysis between TRPVs and clinical traits showed that TRPVs were related to pathologic stage, TNM stage and first course treatment outcome. Moreover, TRPV expression was highly correlated with MSI and TMB. Immunotherapy is a research hotspot at present, our result showed the significant association between TRPVs expression and immune infiltration indicated that TRPV expression alternation could be used to guide prognosis. In addition, we also discovered that the expression level of TRPV1/2/3/4/6 was positively or negatively correlated with objective responses to anti-PD-1/PD-L1 across multiple immunotherapy cohort. Further analysis of drug sensitivity showed the value to treatment. Based on the above analysis, we next focused on TRPV family in bladder cancer. The result demonstrated TRPV also played an important role in bladder cancer. Finally, qPCR assay verified our analysis in bladder cancer. Conclusion: Our study firstly revealed expression and genome alternation of TRPV in pan-cancer. TRPV could be used to predict prognosis or instructing treatment of human cancers, especially bladder cancer.

As a no-disturbance integrated-retrofitting technique, an external rocking frame was widely used on reinforced concrete (RC) structures. Yet, with the increasing demand for seismic strengthening of existing buildings, it has become a concern to evaluate the seismic strengthening schemes based on seismic resilience. Firstly, the dynamic equation of the structural system was derived, and the deformation control mechanism was revealed; thus, the corresponding design method was put forward for the rocking frame reinforcement. Secondly, after soft-first-floor structures were reinforced by rocking frames, the evaluation method of the reinforcement scheme was investigated based on seismic resilience. Finally, the feasibility of the assessment method was verified by a soft-first-floor frame structure, and a comparison was made between the method proposed in this paper and the conventional method. The results find that the soft-first-floor structure reinforced by the rocking frame increased by 10% in the inter-layer displacement and improved by 55.6% and 63.0% in the injury and mortality rates, compared to the buckling-restrained brace scheme. This indicates that the reinforcement scheme of soft layer structures with rocking frames is feasible and effective, and the reinforcement evaluation method proposed in this paper can quantitatively reflect the improvement in seismic performance.

Semantic segmentation of Signet Ring Cells (SRC) plays a pivotal role in the diagnosis of SRC carcinoma based on pathological images. Deep learning-based methods have demonstrated significant promise in computer-aided diagnosis over the past decade. However, many existing approaches rely heavily on stacking layers, leading to repetitive computational tasks and unnecessarily large neural networks. Moreover, the lack of available ground truth data for SRCs hampers the advancement of segmentation techniques for these cells. In response, this paper introduces an efficient and accurate deep learning framework (RGGC-UNet), which is a UNet framework including our proposed residual ghost block with ghost coordinate attention, featuring an encoder-decoder structure tailored for the semantic segmentation of SRCs. We designed a novel encoder using the residual ghost block with proposed ghost coordinate attention. Benefiting from the utilization of ghost block and ghost coordinate attention in the encoder, the computational overhead of our model is effectively minimized. For practical application in pathological diagnosis, we have enriched the DigestPath 2019 dataset with fully annotated mask labels of SRCs. Experimental outcomes underscore that our proposed model significantly surpasses other leading-edge models in segmentation accuracy while ensuring computational efficiency.

This paper proposes a fast direction of arrival (DOA) estimation method based on positive incremental modified Cholesky decomposition atomic norm minimization (PI-CANM) for augmented coprime array sensors. The approach incorporates coprime sampling on the augmented array to generate a non-uniform, discontinuous virtual array. It then utilizes interpolation to convert this into a uniform, continuous virtual array. Based on this, the problem of DOA estimation is equivalently formulated as a gridless optimization problem, which is solved via atomic norm minimization to reconstruct a Hermitian Toeplitz covariance matrix. Furthermore, by positive incremental modified Cholesky decomposition, the covariance matrix is transformed from positive semi-definite to positive definite, which simplifies the constraint of optimization problem and reduces the complexity of the solution. Finally, the Multiple Signal Classification method is utilized to carry out statistical signal processing on the reconstructed covariance matrix, yielding initial DOA angle estimates. Experimental outcomes highlight that the PI-CANM algorithm surpasses other algorithms in estimation accuracy, demonstrating stability in difficult circumstances such as low signal-to-noise ratios and limited snapshots. Additionally, it boasts an impressive computational speed. This method enhances both the accuracy and computational efficiency of DOA estimation, showing potential for broad applicability.

Investigating the impacts of climate change on soil desiccation in planted forests is of theoretical and practical importance for water resource management and ecological restoration in the Loess Plateau. In this study, field experiments were conducted in four planted forests (Platycladus orientalis plots) with tree ages of 5, 12, 25, and 35a (abbreviated as T5, T12, T25, and T35 plots, respectively). Based on which, Hydrus–1D models were calibrated and validated to simulate soil water dynamics under preset climate change scenarios. Then, the soil desiccation index (SDI) in different soil layers was calculated and evaluated. The results showed that under historical climate scenarios, soil desiccation levels were slight or medium in T5, T12, and T25 plots but were mainly medium or severe in T35 plots. Future warming and drying climate would greatly increase the SDI in all plots, and its impacts increased with tree ages in the upper soil layers (0–40 cm) but decreased with it in lower layers (40–100 cm). However, it only had limited influence on the responses of soil desiccation to the climate in different hydrological years. Precipitation and vapour pressure deficit dominated soil desiccation in the upper and lower layers, respectively, and both showed linear relationships (p

Background: This study constructs a molecular subtype and prognostic model of bladder cancer (BLCA) through endoplasmic reticulum stress (ERS) related genes, thus helping to clinically guide accurate treatment and prognostic assessment.Methods: The Bladder Cancer (BLCA) gene expression data was downloaded from The Cancer Genome Atlas (TCGA) and Gene Expression Omnibus (GEO) database. We clustered by ERS-related genes which obtained through GeneCards database, results in the establishment of a new molecular typing of bladder cancer. Further, we explored the characteristics of each typology in terms of immune microenvironment, mutations, and drug screening. By analyzing the ERS-related genes with univariate Cox, LASSO and multivariate Cox analyses, we also developed the four-gene signature, while validating the prognostic effect of the model in GSE32894 and GSE13507 cohorts. Finally, we evaluated the prognostic value of the clinical data in the high and low ERS score groups and constructed a prognostic score line graph by Nomogram.Results: We constructed four molecular subtypes (C1- C4) of bladder cancer, in which patients with C2 had a poor prognosis and those with C3 had a better prognosis. The C2 had a high degree of TP53 mutation, significant immune cell infiltration and high immune score. In contrast, C3 had a high degree of FGFR3 mutation, insignificant immune cell infiltration, and reduced immune checkpoint expression. After that, we built ERS-related risk signature to calculate ERS score, including ATP2A3, STIM2, VWF and P4HB. In the GSE32894 and GSE13507, the signature also had good predictive value for prognosis. In addition, ERS scores were shown to correlate well with various clinical features. Finally, we correlated the ERS clusters and ERS score. Patients with high ERS score were more likely to have the C2 phenotype, while patients with low ERS score were C3.Conclusion: In summary, we identified four novel molecular subtypes of BLCA by ERS-related genes which could provide some new insights into precision medicine. Prognostic models constructed from ERS-related genes can be used to predict clinical outcomes. Our study contributes to the study of personalized treatment and mechanisms of BLCA.

ObjectiveThe association between serum uric acid (UA) and spontaneous hemorrhagic transformation (HT) has been seldom studied, and the role of UA in spontaneous HT remains unclear. This study aims to investigate the sex-dependent association between UA and spontaneous HT in patients with ischemic stroke.MethodWe retrospectively included patients with ischemic stroke in a tertiary academic hospital between December 2016 and May 2020. Patients were included if they presented within 24 h after the onset of symptoms and did not receive reperfusion therapy. Spontaneous HT was determined by an independent evaluation of neuroimaging by three trained neurologists who were blinded to clinical data. A univariate analysis was performed to identify factors related to spontaneous HT. Four logistic regression models were established to adjust each factor and assess the association between UA and spontaneous HT.ResultsA total of 769 patients were enrolled (64.6% were male patients and 3.9% had HT). After adjusting the confounders with a P < 0.05 (model A) in the univariate analysis, the ratio of UA and its interquartile range (RUI) was independently associated with spontaneous HT in male patients (OR: 1.85; 95% CI: 1.07–3.19; P = 0.028), but not in female patients (OR: 1.39; 95% CI: 0.28–6.82; P = 0.685). In models B–D, the results remain consistent with model A after the adjustment for other potential confounders.ConclusionsHigher serum UA was independently associated with a higher occurrence of spontaneous HT in male patients who were admitted within 24 h after the stroke onset without receiving reperfusion therapy.

Colorectal cancer (CRC) is a prevalent gastrointestinal tumour with high incidence and mortality rates. Early screening for CRC can improve cure rates and reduce mortality. Recently, deep convolution neural network (CNN)-based pathological image diagnosis has been intensively studied to meet the challenge of time-consuming and labour-intense manual analysis of high-resolution whole slide images (WSIs). Despite the achievements made, deep CNN-based methods still suffer from some limitations, and the fundamental problem is that they cannot capture global features. To address this issue, we propose a hybrid deep learning framework (RGSB-UNet) for automatic tumour segmentation in WSIs. The framework adopts a UNet architecture that consists of the newly-designed residual ghost block with switchable normalization (RGS) and the bottleneck transformer (BoT) for downsampling to extract refined features, and the transposed convolution and 1 × 1 convolution with ReLU for upsampling to restore the feature map resolution to that of the original image. The proposed framework combines the advantages of the spatial-local correlation of CNNs and the long-distance feature dependencies of BoT, ensuring its capacity of extracting more refined features and robustness to varying batch sizes. Additionally, we consider a class-wise dice loss (CDL) function to train the segmentation network. The proposed network achieves state-of-the-art segmentation performance under small batch sizes. Experimental results on DigestPath2019 and GlaS datasets demonstrate that our proposed model produces superior evaluation scores and state-of-the-art segmentation results.

Using SnCl2·2H2O and SnCl4·5H2O as tin source and polyvinylpyrrolidone (PVP) as polymer matrix, fibril was prepared by using simple and controllable electrostatic spinning technology, and SnO2 nanofibers were obtained after calcination. Scanning electron microscope (SEM), X-ray diffractometer (XRD) etc. were used to investigate the effect of PVP molecular weight and different tin sources on the morphological properties of SnO2 nanofibers, and the specific surface area and gas-sensing characteristics of SnO2 nanofibers were tested. The results show that the fiber precursor prepared by electrospinning can remove the polymer matrix and solvent after calcination, and obtain pure tetragonal rutile SnO2 nanofibers, which show the characteristics of large specific surface area, high sensitivity and fast response speed. When PVP-K30 is used as high molecular matrix and SnCl4 · 5H2O is used as tin source, the prepared SnO2 nanofibers have uniform diameter, clear fiber, good orientation and better gas sensing characteristics.

High-fat diet could lead to a series of metabolic diseases, including obesity, and its mechanism is not clear. In this study, the rabbit individuals were fed with high-fat diet, the liver tissues were collected, high-throughput sequencing technology was used to reveal the expression of lncRNA and miRNA difference, and the molecular regulation mechanism of lncRNA-miRNA. A total of 24,615 DE lncRNAs and 52 DE miRNAs were identified, including 15 novel discovered DE miRNAs (5 upregulated and 10 downregulated). Furthermore, five miRNAs and three mRNAs were verified by qRT-PCR, and the results showed that the expression of the DE miRNAs and DE lncRNAs in the two groups was consistent with our sequencing results. GO and KEGG analyzed 7,57,139 target genes respectively, enriching the pathways related to lipid metabolism, including mucin O-glycan biosynthesis pathway, insulin resistance and glucagon signaling pathway. Moreover, 65 targeting relationships were obtained. Among them, LOC103348122/miR-450a-5p, LOC103350359/miR-450a-3p and LOC103350429/miR-148a-5p were proposed the first time. Significantly, LOC103348122/miR-450a-5p and LOC103350429/miR-148a-5p were related to lipid metabolism in the liver. This study is of great significance to the CeRNA regulatory network related to lipid metabolism in the liver of rabbits, and provides a basis for understanding hepatic steatosis in rabbits.

As an important technology in 3D vision, point-cloud registration has broad development prospects in the fields of space-based remote sensing, photogrammetry, robotics, and so on. Of the available algorithms, the Iterative Closest Point (ICP) algorithm has been used as the classic algorithm for solving point cloud registration. However, with the point cloud data being under the influence of noise, outliers, overlapping values, and other issues, the performance of the ICP algorithm will be affected to varying degrees. This paper proposes a global structure and adaptive weight aware ICP algorithm (GSAW-ICP) for image registration. Specifically, we first proposed a global structure mathematical model based on the reconstruction of local surfaces using both the rotation of normal vectors and the change in curvature, so as to better describe the deformation of the object. The model was optimized for the convergence strategy, so that it had a wider convergence domain and a better convergence effect than either of the original point-to-point or point-to-point constrained models. Secondly, for outliers and overlapping values, the GSAW-ICP algorithm was able to assign appropriate weights, so as to optimize both the noise and outlier interference of the overall system. Our proposed algorithm was extensively tested on noisy, anomalous, and real datasets, and the proposed method was proven to have a better performance than other state-of-the-art algorithms.

Recently, deep learning methods have widely been employed for surface defect segmentation in industrial production with remarkable success. Nevertheless, accurate segmentation of various types of defects is still challenging due to their irregular appearance and low contrast with the background. In light of this challenge, we propose an attention-based network with a U-shaped structure, referred to as AFFNet. In the encoder part, we present a newly designed module, Residual-RepGhost-Dblock (RRD), which focuses on the extraction of more representative features using CA attention and dilated convolution with varying expansion rates without a concomitant increase in the parameters. In the decoder part, we introduce a novel global feature attention (GFA) module to selectively fuse low-level and high-level features, suppressing distracting information such as background. Moreover, considering the imbalance of the dataset sampled from actual industrial production and the difficulty of training samples with small defects, we use the online hard sample mining (OHEM) cross-entropy loss function to improve the learning ability of hard samples. Experimental results on the NEU-seg dataset demonstrate the superiority of our method over other state-of-the-art methods.

This paper proposes a human activity recognition (HAR) method for frequency-modulated continuous wave (FMCW) radar sensors. The method utilizes a multi-domain feature attention fusion network (MFAFN) model that addresses the limitation of relying on a single range or velocity feature to describe human activity. Specifically, the network fuses time-Doppler (TD) and time-range (TR) maps of human activities, resulting in a more comprehensive representation of the activities being performed. In the feature fusion phase, the multi-feature attention fusion module (MAFM) combines features of different depth levels by introducing a channel attention mechanism. Additionally, a multi-classification focus loss (MFL) function is applied to classify confusable samples. The experimental results demonstrate that the proposed method achieves 97.58% recognition accuracy on the dataset provided by the University of Glasgow, UK. Compared to existing HAR methods for the same dataset, the proposed method showed an improvement of about 0.9–5.5%, especially in the classification of confusable activities, showing an improvement of up to 18.33%.

Sketch face recognition refers to the process of matching sketches to photos. Recently, there has been a growing interest in using deep learning to learn discriminative features for sketch face recognition. However, the success of deep learning relies on the large-scale paired images to counteract effects such as over-fitting, since the amount of the paired training data is relatively small, the discriminative power of the deeply learned features will inevitably be reduced. This paper proposes a novel deep metric learning method termed domain alignment embedding network for sketch face recognition. Specifically, a training episode strategy is designed to alleviate the small sample problem, and a domain alignment embedding loss is proposed to guide the feature embedding network to learn discriminative features. Extensive experimental results on the UoM-SGFSv2 and PRIP-VSGC datasets are verified to show the effectiveness of the proposed method.

Study region: A typical grassland (Medicago sativa plot) and a typical forest land (Robinia pseudoacacia plot) located in the semi-arid Loess Plateau, China. Study focus: Field observations and scenario simulation approaches were used to investigate the potential impact of future warming and drying climate on the whole water budget and to assess whether, how, and to what degree the highly recommended man–driven tree–to–grass conversion could alleviate the adverse impact of such climate change on soil water. New hydrological insights for the region: Differences of interception and transpiration between two plots resulted in their different soil water storage. Under the warming and drying future climate, interception and transpiration in two plots would experience significant reductions, especially transpiration, and M. sativa plot could maintain more water than R. pseudoacacia plot. When R. pseudoacacia was replaced by M. sativa, although interception would dramatically increase, the transpiration would significantly decrease, which made the soil water storage experience an insignificant change. Climate change and tree–to–grass conversion mainly exerted their influences in months with larger precipitation, and vegetation conversion would reduce monthly T/ET. Our study improves the understanding of the hydrological processes under climate change and vegetation conversion, and also provides valuable guidelines for the management of water resources and the restoration of ecological environment in the semi–arid Loess Plateau.

In recent years, deep learning has achieved good results in the semantic segmentation of medical images. A typical architecture for segmentation networks is an encoder–decoder structure. However, the design of the segmentation networks is fragmented and lacks a mathematical explanation. Consequently, segmentation networks are inefficient and less generalizable across different organs. To solve these problems, we reconstructed the segmentation network based on mathematical methods. We introduced the dynamical systems view into semantic segmentation and proposed a novel segmentation network based on Runge–Kutta methods, referred to hereafter as the Runge–Kutta segmentation network (RKSeg). RKSegs were evaluated on ten organ image datasets from the Medical Segmentation Decathlon. The experimental results show that RKSegs far outperform other segmentation networks. RKSegs use few parameters and short inference time, yet they can achieve competitive or even better segmentation results compared to other models. RKSegs pioneer a new architectural design pattern for segmentation networks.

Nuclei segmentation and classification are two basic and essential tasks in computer-aided diagnosis of digital pathology images, and those deep-learning-based methods have achieved significant success. Unfortunately, most of the existing studies accomplish the two tasks by splicing two related neural networks directly, resulting in repetitive computation efforts and a redundant-and-large neural network. Thus, this paper proposes a lightweight deep learning framework (GSN-HVNET) with an encoder–decoder structure for simultaneous segmentation and classification of nuclei. The decoder consists of three branches outputting the semantic segmentation of nuclei, the horizontal and vertical (HV) distances of nuclei pixels to their mass centers, and the class of each nucleus, respectively. The instance segmentation results are obtained by combing the outputs of the first and second branches. To reduce the computational cost and improve the network stability under small batch sizes, we propose two newly designed blocks, Residual-Ghost-SN (RGS) and Dense-Ghost-SN (DGS). Furthermore, considering the practical usage in pathological diagnosis, we redefine the classification principle of the CoNSeP dataset. Experimental results demonstrate that the proposed model outperforms other state-of-the-art models in terms of segmentation and classification accuracy by a significant margin while maintaining high computational efficiency.

A three-dimensional finite element analysis was carried out for the influence of soil-structure interaction (SSI) on the seismic response of an isolated continuous bridge with a friction pendulum system (FPS). SSI was modelled using the soil resistance vs deflection nonlinear springs (following p-y curves). The influence of SSI on the dynamic behaviour and the seismic response of isolated bridges with different friction coefficients of FPS were investigated using six measured ground motions. Results show that the effect of SSI is limited on the isolation period of the bridge, but significant on the response amplitude of the bridge structures, and particularly on the response amplitude at the pier top and bottom. The displacement and acceleration at the pier top are underestimated, while the shear force and bending moment at the pier bottom are overestimated if the SSI effects are ignored, which suggests that SSI should be considered in the seismic analyses of continuous bridges.

Cervical cancer is one of the most common cancers that threaten women’s lives, and its early screening is of great significance for the prevention and treatment of cervical diseases. Pathologically, the accurate segmentation of cervical cells plays a crucial role in the diagnosis of cervical cancer. However, the frequent presence of adherent or overlapping cervical cells in Pap smear images makes separating them individually a difficult task. Currently, there are few studies on the segmentation of adherent cervical cells, and the existing methods commonly suffer from low segmentation accuracy and complex design processes. To address the above problems, we propose a novel star-convex polygon-based convolutional neural network with an encoder-decoder structure, called SPCNet. The model accomplishes the segmentation of adherent cells relying on three steps: automatic feature extraction, star-convex polygon detection, and non-maximal suppression (NMS). Concretely, a new residual-based attentional embedding (RAE) block is suggested for image feature extraction. It fuses the deep features from the attention-based convolutional layers with the shallow features from the original image through the residual connection, enhancing the network’s ability to extract the abundant image features. And then, a polygon-based adaptive NMS (PA-NMS) algorithm is adopted to screen the generated polygon proposals and further achieve the accurate detection of adherent cells, thus allowing the network to completely segment the cell instances in Pap smear images. Finally, the effectiveness of our method is evaluated on three independent datasets. Extensive experimental results demonstrate that the method obtains superior segmentation performance compared to other well-established algorithms.

In intelligent traffic monitoring, speed measuring millimeter waves (MMW) radar is one of the most commonly used tools for traffic enforcement. In traffic enforcement field, the radar must provide the evidence of each vehicle belongs to which lane. In this paper, we propose a novel kernel line segment adaptive possibilistic c-means clustering algorithm (KLSAPCM) for lane determination of vehicles. Firstly, the raw measurement data is preprocessed using the extracting method of data adjacent lane centerlines. Secondly, according to the improved minimum radius data search method, outliers are removed and the proposed KLSAPCM algorithm is initialized. Finally, the accuracy of lane determination has been improved by the proposed KLSAPCM clustering algorithm based on adaptive kernel line segment that conforms to the shape features of the measurement data in the actual scene. The experiment results for multiple scenes were: the KLSAPCM algorithm is compared with the DBSCAN, the $k$ -means, the FCM, the PCM, the AMPCM, and the APCM algorithms on real measurement datasets, and the results highlight the classification rate of the proposed algorithm. Meanwhile, the proposed algorithm gets a good real-time performance and strong robustness for some sparse moving vehicle scene applications.

Nowadays, faces in videos can be easily replaced with the development of deep learning, and these manipulated videos are realistic and cannot be distinguished by human eyes. Some people maliciously use the technology to attack others, especially celebrities and politicians, causing destructive social impacts. Therefore, it is imperative to design an accurate method for detecting face manipulation. However, most of the existing methods adopt single convolutional neural network as the feature extraction module, causing the extracted features to be inconsistent with the human visual mechanism. Moreover, the rich details and semantic information cannot be reflected with single feature, limiting the detection performance. Therefore, this paper tackles the above problems by proposing a novel face manipulation detection method based on a supervised multi-feature fusion attention network (SMFAN). Specifically, the capsule network is used for face manipulation detection, and the SMFAN is added to the original capsule network to extract details of the fake face image. Further, the focal loss is used to realize hard example mining. Finally, the experimental results on the public dataset FaceForensics++ show that the proposed method has better performance.

Aimed at the problems in which the performance of filters derived from a hypothetical model will decline or the cost of time of the filters derived from a posterior model will increase when prior knowledge and second-order statistics of noise are uncertain, a new filter is proposed. In this paper, a Bayesian robust Kalman filter based on posterior noise statistics (KFPNS) is derived, and the recursive equations of this filter are very similar to that of the classical algorithm. Note that the posterior noise distributions are approximated by overdispersed black-box variational inference (O-BBVI). More precisely, we introduce an overdispersed distribution to push more probability density to the tails of variational distribution and incorporated the idea of importance sampling into two strategies of control variates and Rao–Blackwellization in order to reduce the variance of estimators. As a result, the convergence process will speed up. From the simulations, we can observe that the proposed filter has good performance for the model with uncertain noise. Moreover, we verify the proposed algorithm by using a practical multiple-input multiple-output (MIMO) radar system.

In the multi-target traffic radar scene, the clustering accuracy between vehicles with close driving distance is relatively low. In response to this problem, this paper proposes a new clustering algorithm, namely an adaptive ellipse distance density peak fuzzy (AEDDPF) clustering algorithm. Firstly, the Euclidean distance is replaced by adaptive ellipse distance, which can more accurately describe the structure of data obtained by radar measurement vehicles. Secondly, the adaptive exponential function curve is introduced in the decision graph of the fast density peak search algorithm to accurately select the density peak point, and the initialization of the AEDDPF algorithm is completed. Finally, the membership matrix and the clustering center are calculated through successive iterations to obtain the clustering result.The time complexity of the AEDDPF algorithm is analyzed. Compared with the density-based spatial clustering of applications with noise (DBSCAN), k-means, fuzzy c-means (FCM), Gustafson-Kessel (GK), and adaptive Euclidean distance density peak fuzzy (Euclid-ADDPF) algorithms, the AEDDPF algorithm has higher clustering accuracy for real measurement data sets in certain scenarios. The experimental results also prove that the proposed algorithm has a better clustering effect in some close-range vehicle scene applications. The generalization ability of the proposed AEDDPF algorithm applied to other types of data is also analyzed.