Search

Abstract Weakly supervised semantic segmentation (WSSS) is a challenging task of computer vision. The state-of-the-art semantic segmentation methods are usually based on the convolutional neural network (CNN), which mainly have the drawbacks of inability to explore the global information correctly and failure to activate potential object regions. To avoid such drawbacks, the transformer approach is explored in the WSSS task, but no effective semantic association between different patch tokens can be determined in the transformer. To address this issue, inspired by the graph convolutional network (GCN), this paper proposes a graph structure to learn the semantic category relationships between different blocks in the vector sequence. To verify the effectiveness of the proposed method in this paper, a large number of experiments were conducted on the publicly available PASCAL VOC2012 dataset. The experimental results show that our proposed method achieves significant performance improvement in the WSSS task and outperforms other state-of-the-art transformer-based methods.

The coronavirus disease 2019 (COVID-19) outbreak has resulted in countless infections and deaths worldwide, posing increasing challenges for the health care system. The use of artificial intelligence to assist in diagnosis not only had a high accuracy rate but also saved time and effort in the sudden outbreak phase with the lack of doctors and medical equipment. This study aimed to propose a weakly supervised COVID-19 classification network (W-COVNet). This network was divided into three main modules: weakly supervised feature selection module (W-FS), deep learning bilinear feature fusion module (DBFF) and Grad-CAM++ based network visualization module (Grad-Ⅴ). The first module, W-FS, mainly removed redundant background features from computed tomography (CT) images, performed feature selection and retained core feature regions. The second module, DBFF, mainly used two symmetric networks to extract different features and thus obtain rich complementary features. The third module, Grad-Ⅴ, allowed the visualization of lesions in unlabeled images. A fivefold cross-validation experiment showed an average classification accuracy of 85.3%, and a comparison with seven advanced classification models showed that our proposed network had a better performance.

Summary: This paper proposes a novel clustering and dynamic recognition–based auto-reservoir neural network (CDbARNN) for short-term load forecasting (STLF) of industrial park microgrids. In CDbARNN, the available load sets are first decomposed into several clusters via K-means clustering. Then, by extracting characteristic information of the load series input to CDbARNN and the load curves belonging to each cluster center, a dynamic recognition technology is developed to identify which cluster of the input load series belongs to. After that, the input load series and the load curves of the cluster to which it belongs constitute a short-term high-dimensional matrix entered into the reservoir of CDbARNN. Finally, reservoir node numbers of CDbARNN which are used to match different clusters are optimized. Numerical experiments conducted on STLF of an actual industrial park microgrid indicate the dominating performance of the proposed approach through several cases and comparisons with other well-known deep learning methods.

The electrocatalytic nitrogen reduction reaction (NRR) is considered a viable alternative to the Haber–Bosch process for ammonia synthesis, and the design of highly active and selective catalysts is crucial for the industrialization of the NRR. Dual-atom catalysts (DACs) with dual active sites offer flexible active sites and synergistic effects between atoms, providing more possibilities for the tuning of catalytic performance. In this study, we designed 48 graphene-based DACs with N4O2 coordination (MM′@N4O2-G) using density functional theory. Through a series of screening strategies, we explored the reaction mechanisms of the NRR for eight catalysts in depth and revealed the “acceptance–donation” mechanism between the active sites and the N2 molecules through electronic structure analysis. The study found that the limiting potential of the catalysts exhibited a volcano-shaped relationship with the d-band center of the active sites, indicating that the synergistic effect between the bimetallic components can regulate the d-band center position of the active metal M, thereby controlling the reaction activity. Furthermore, we investigated the selectivity of the eight DACs and identified five potential NRR catalysts. Among them, MoCo@N4O2-G showed the best NRR performance, with a limiting potential of −0.20 V. This study provides theoretical insights for the design and development of efficient NRR electrocatalysts.

In order to further study the expansion characteristics of left-turning non-motorized vehicles at intersections and the relationship between expansion characteristics and vehicle-bicycle conflicts, the trajectory point data of left-turning non-motorized vehicles are extracted using video trajectory tracking technology, and construct the cubic curve expansion envelope equation with the highest fitting degree. For the purpose of quantifying the expansion degree of non-motor vehicles after starting, two intersections in Guangxi Zhuang Autonomous Region were selected for case analysis, and the numerical range of expansion degree of the intersection with a left-turn waiting area and the intersection without a left-turn waiting area was obtained. Study the mathematical relationship between the expansion degree and its influencing factors, and establish the multivariate nonlinear regression equation between the expansion degree and the left-turn non-motorized vehicle flow, the number of parallel non-motorized vehicles, and the left-turn green light time. Analyze the vehicle-bicycle conflicts caused by the expansion of left-turning non-motorized vehicles, determine the essential factors affecting the number of non-motorized vehicles, and establish the multiple linear regression equation between the number of non-motorized vehicles and the number of left-turning non-motorized vehicles, the expansion degree, and the number of parallel non-motorized vehicles, the results show that the model has high accuracy. By analyzing the expansion characteristics of left-turning non-motorized vehicles at intersections, the relationship between different influencing factors and the expansion degree is obtained. Then the vehicle-bicycle conflicts under the influence of expansion characteristics is analyzed, providing theoretical ideas for improving traffic efficiency and optimizing traffic organization at intersections.

IntroductionChronic disease self-management is influenced by many factors. Previous studies have linked patients’ media use with chronic disease self-management, but the underlying mechanisms of this relationship are less understood.ObjectivesThe purpose of this study is to explore the mediating role of family health (FH) between frequency of smartphone use (FOSU) and self-management behaviors among middle-aged and older patients with chronic diseases (SBAMAOPWCD) through a moderated mediation model, and whether this indirect relationship is modified by the solitary status of middle-aged and older Chinese patients with chronic disease.MethodsSurveys were collected from 1,424 (N = 1,424; age > 45) middle-aged and older with one or more chronic conditions in China on self-reports of FOSU, FH and Chronic disease self-management behaviors were used to examine the moderated mediation model.ResultsThe results showed that the FOSU was significantly and positively associated with SBAMAOPWCD (β = 0.220, p

Abstract The Internet of things for environment monitoring requires high performance with low power-consumption gas sensors which could be easily integrated into large-scale sensor network. While semiconductor gas sensors have many advantages such as excellent sensitivity and low cost, their application is limited by their high operating temperature. Two-dimensional (2D) layered materials, typically molybdenum disulfide (MoS2) nanosheets, are emerging as promising gas-sensing materials candidates owing to their abundant edge sites and high in-plane carrier mobility. This work aims to overcome the sluggish and weak response as well as incomplete recovery of MoS2 gas sensors at room temperature by sensitizing MoS2 nanosheets with PbS quantum dots (QDs). The huge amount of surface dangling bonds of QDs enables them to be ideal receptors for gas molecules. The sensitized MoS2 gas sensor exhibited fast and recoverable response when operated at room temperature, and the limit of NO2 detection was estimated to be 94 ppb. The strategy of sensitizing 2D nanosheets with sensitive QD receptors may enhance receptor and transducer functions as well as the utility factor that determine the sensor performance, offering a powerful new degree of freedom to the surface and interface engineering of semiconductor gas sensors.

For the discrimination of false targets, the discrimination probability can be improved by increasing the number of radar stations. However, that may result in a serious waste of equipment resources when too many radars are involved. An asymptotic subset selection strategy based on target positioning characteristics is proposed to address the above issues. Several effective strategies are considered to select some transmitters and receivers to form a radar subset, such as the rapid shrinkage method, global shrinkage method, and predetermined size method, which can guarantee the preset discrimination performance of limited equipment resources and reduce the waste of resources. All of the selected stations have good spatial distribution or strong discrimination capacity in multistatic radar system. Compared with the exhaustive search, the proposed subset selection strategy affords a significant reduction in terms of time complexity. The simulation results show that the radar subset can maintain approximate discrimination performance with the original multistatic radar systems. At the same time, the proposed method optimizes the number of radar stations and reduces data processing time and required communication links, thus effectively saving operating costs.

Activity of human CYP26A1 towards six proluciferin probe substrates and their ester derivatives was monitored. These included three monofluorobenzyl ether isomers and three five-membered heterocycles. Overall, luciferin substrates with a free acid group gave higher activities than the ester compounds. Also, luciferin derivatives with six-ring structures were better metabolized than those with five-rings. The best substrates identified in this study are Luciferin 6′ 3-fluorobenzyl ether (Luciferin-3FBE) and its methyl ester (Luciferin-3FBEME). Taken together, we describe eleven new probe substrates for CYP26A1 and demonstrate for the first time that CYP26A1 does not only accept acid substrates but can also metabolize esters.

Semiconducting materials are potential SERS substrates as alternatives to noble metals, but often suffer from poor stabilities and sensitivities. Here, the authors use molybdenum dioxide as a SERS material, showing high enhancement factors and stability to oxidation even at high temperatures.

Abstract Background Adrenomyeloneuropathy (AMN) is a rare genetic disease. In this study, a case of AMN was uncovered in a Chinese family. Methods Clinical manifestations were collected and observed through medical records, physical examination, laboratory tests, and magnetic resonance imaging (MRI). Generation sequencing of the ABCD1 gene was performed, and the pedigree of the family was analyzed. Results The proband suffered from adrenocortical insufficiency at 8 years old and presented with a slowly progressive gait disorder at 21 years old. Physical examination, laboratory tests, and MRI showed that he had adult‐onset AMN manifestations, including spasticity and hyperactive tendon reflexes with Hoffman and Babinski signs in the limbs, difficulty in performing the heel‐to‐shin test, hyperpigmentation, increased levels of adrenocorticotropic hormone and very long‐chain fatty acids, decreased levels of corticosteroid and serum gesterol, and salient atrophy of the cervical and thoracic spinal cord. DNA analysis revealed a missense variant, c.290A>C (p.His97Pro) in exon 1 of the ABCD1 gene, in the proband. Sanger sequencing confirmed that the proband's mother was heterozygous for the same variant. The ABCD1 gene mutation transmitted in an X‐linked inheritance manner. Conclusion A novel missense mutation in the ABCD1 gene was identified in a Chinese family, which caused an unusual manifestation of adult‐onset AMN. This discovery is beneficial for the genetic counseling of patients with X‐linked adrenoleukodystrophy.

A major feature of acute lung injury (ALI) is excessive inflammation in the lung. Vitexin is an active component from medicinal plants which has antioxidant and anti-inflammatory activities. Oxidative stress and inflammation play important roles in the pathophysiological processes in ALI. In the current study, we investigate the effect and potential mechanisms of Vitexin on lipopolysaccharide (LPS)-induced ALI.ALI was induced by LPS intratracheal instillation in C57BL/6 wild-type mice and Nrf2 gene knocked down (Nrf2-/-) mice. One hour before LPS challenge, Vitexin or vehicle intraperitoneal injection was performed. Bronchoalveolar lavage fluid and lung tissues were examined for lung inflammation and injury at 24 h after LPS challenge.Our animal study's results showed that LPS-induced recruitment of neutrophils and elevation of proinflammatory cytokine levels were attenuated by Vitexin treatment. Vitexin decreased lung edema and alveolar protein content. Moreover, Vitexin activated nuclear factor erythroid-2-related factor 2 (Nrf2), and increased the activity of its target gene heme oxygenase (HO)-1. The LPS-induced reactive oxygen species were inhibited by Vitexin. In addition, the activation of the nucleotide-binding domain and leucine-rich repeat PYD-containing protein 3 (NLRP3) inflammasome was suppressed by Vitexin. However, these effects of Vitexin were abolished in the Nrf2-/- mice. Our cell studies showed that Vitexin enhanced the expression of Nrf2 and HO-1 activity. Moreover, reactive oxygen species (ROS) and IL-1β productions were reduced in Vitexin-treated cells. However, knockdown of Nrf2 by siRNA in RAW cells reversed the benefit of Vitexin.Vitexin suppresses LPS-induced ALI by controlling Nrf2 pathway.

Objective. This study aimed to examine the potential role of memory T follicular helper (Tfh) cells in patients with neuromyelitis optica/neuromyelitis optica spectrum disorders (NMO/NMOSD). Methods. The percentages of different subsets of circulating memory Tfh cells in 25 NMO/NMOSD patients before and after treatment as well as in 17 healthy controls were examined by flow cytometry. The levels of IL-21 and AQP4 Ab in plasma and CSF were measured by ELISA. Results. The percentages and numbers of circulating memory Tfh cells, ICOS+, CCR7−, CCR7−ICOS+, CCR7+, CCR7+ICOS+ memory Tfh cells, and the levels of IL-21 in plasma and CSF were significantly increased in NMO/NMOSD patients. The percentages of CCR7− and CCR7−ICOS+ memory Tfh cells were positively correlated with ARR, plasma IL-21, and AQP4 Ab levels. The percentages of CCR7+ and CCR7+ICOS+ memory Tfh cells were positively correlated with CSF white blood cell counts, proteins, and IL-21 levels. Treatment with corticosteroids significantly reduced the numbers of CCR7−ICOS+ and CCR7+ICOS+ memory Tfh cells as well as plasma IL-21 levels in patients with partial remission. Conclusions. Our findings indicate that circulating memory Tfh cells may participate in the relapse and development of NMO/NMOSD and may serve as a new therapeutic target.

The mechanism for the reaction of radical FOO with NO was investigated theoretically. The electronic structure information for the singlet potential energy surfaces (PES) was obtained at the MP2/6-311+G(2df) level of theory, and the single-point energies were refined by the CCSD(T)/6-311+G(2df) level. The rate constants as well as the pressure and temperature dependence of various product channels for the reaction are predicted. The calculated results show that starting from the energy-rich intermediate FOONOtp (IM1), at room temperature 298 K and at 1 Torr, FNO is the exclusive product which is consistent with the experimental results. The present results will be useful to gain a deep insight into the reaction mechanism of FOO + NO reaction.

Angiostrongylus cantonensis is the most common cause of eosinophilic meningitis in humans. It is usually caused by ingestion of raw or inadequately cooked intermediate hosts or food contaminated with infective third-stage larvae. We describe a case of eosinophilic meningitis caused by A. cantonensis in a male Chinese patient. The patient had a history of eating raw fish and snail. We describe the clinical features of the patient, the diagnostic process and treatments. We also provide a brief update for physicians on the characteristics, diagnosis and treatment of eosinophilic meningitis caused by A. cantonensis, with particular emphasis on the update of prevalence and treatment of the disease in China.

The coronavirus disease 2019 (COVID-19) has severely disrupted both human life and the health care system. Timely diagnosis and treatment have become increasingly important; however, the distribution and size of lesions vary widely among individuals, making it challenging to accurately diagnose the disease. This study proposed a deep-learning disease diagnosismodel based onweakly supervised learning and clustering visualization (W_CVNet) that fused classification with segmentation. First, the data were preprocessed. An optimizable weakly supervised segmentation preprocessing method (O-WSSPM) was used to remove redundant data and solve the category imbalance problem. Second, a deep-learning fusion method was used for feature extraction and classification recognition. A dual asymmetric complementary bilinear feature extraction method (D-CBM) was used to fully extract complementary features, which solved the problem of insufficient feature extraction by a single deep learning network. Third, an unsupervised learning method based on Fuzzy C-Means (FCM) clustering was used to segment and visualize COVID-19 lesions enabling physicians to accurately assess lesion distribution and disease severity. In this study, 5-fold cross-validation methods were used, and the results showed that the network had an average classification accuracy of 85.8%, outperforming six recent advanced classification models. W_CVNet can effectively help physicians with automated aid in diagnosis to determine if the disease is present and, in the case of COVID-19 patients, to further predict the area of the lesion. [ABSTRACT FROM AUTHOR]