Search

Dye wastewater discharges threaten fragile aquatic ecosystems. Photocatalysis has been applied as an effective, viable, and economical option to achieve dye degradation. However, current photocatalysts suffer from large bandgaps, powder agglomeration, difficult recovery, and cannot be recycled. Herein, a novel Bi2S3/ZnS heterojunction loaded cellulose/chitosan (BiZnCCS) with highly efficient photocatalytic activities for CR pollutant degradation was designed and fabricated via a simple hydrothermal method. BiZnCCS has a large pore size of 2 μm with a high porosity of 71.62%, which increases the contact area with CR and facilitates the adsorption of CR. The pore wall was uniformly loaded with Bi2S3/ZnS particles, and the degradation rate of CR could reach 97.8% under simulated visible light irradiation. By adjusting the composition ratio, BiZnCCS was able to withstand a stress of about 82.45 kPa, and the CR removal rate was still maintained at about 85% after 7 cycles of testing. By creating a synergistic adsorption-photocatalytic degradation mechanism, BiZnCCS has the ability to efficiently remove CR under visible light, which provides a new material for the treatment of dye wastewater. [ABSTRACT FROM AUTHOR]

Background: Lumbar disc herniation is a common degenerative lumbar disease with an increasing incidence. Percutaneous endoscopic lumbar discectomy can treat lumbar disc herniation safely and effectively with a minimally invasive procedure. However, the learning curve of this technology is steep, which means that initial learners are often not sufficiently proficient in endoscopic operations, which can easily lead to iatrogenic damage. At present, the application of computer deep learning technology to clinical diagnosis, treatment, and surgical navigation has achieved satisfactory results. Purpose: The objective of our team is to develop a multi-element identification system for the visual field of endoscopic spine surgery using deep learning algorithms and to evaluate the feasibility of this system. Method: We established an image database by collecting surgical videos of 48 patients diagnosed with lumbar disc herniation, which was labeled by two spinal surgeons. We selected 6000 images of the visual field of percutaneous endoscopic spine surgery (including various tissue structures and surgical instruments), divided into the training data, validation data, and test data according to 2:1:2. We developed convolutional neural network models based on instance segmentation—Solov2, CondInst, Mask R-CNN and Yolact, and set the four network model backbone as ResNet101 and ResNet50 respectively. Mean average precision (mAP) and frames per second (FPS) were used to measure the performance of each model for classification, localization and recognition in real time, and AP (average) is used to evaluate how easily an element is detected by neural networks based on computer deep learning. Result: Comprehensively comparing mAP and FSP of each model for bounding box test and segmentation task for the test set of images, we found that Solov2 (ResNet101) (mAP = 73.5%, FPS = 28.9), Mask R-CNN (ResNet101) (mAP = 72.8%, FPS = 28.5) models are the most stable, with higher precision and faster image processing speed. Combining the average precision of the elements in the bounding box test and segmentation tasks in each network, the AP(average) was highest for tool 3 (bbox-0.85, segm-0.89) and lowest for tool 5 (bbox-0.63, segm-0.72) in the instrumentation, whereas in the anatomical tissue elements, the fibrosus annulus (bbox-0.68, segm-0.69) and ligamentum flavum (bbox-0.65, segm-0.62) had higher AP(average),while extra-dural fat (bbox-0.42, segm-0.44) was lowest. Conclusion: Our team has developed a multi-element identification system for the visual field of percutaneous endoscopic spine surgery adapted to the interlaminar and foraminal approaches, which can identify and track anatomical tissue (nerve, ligamentum flavum, nucleus pulposus, etc.) and surgical instruments (endoscopic forceps, an high-speed diamond burr, etc.), which can be used in the future as a virtual educational tool or applied to the intraoperative real-time assistance system for spinal endoscopic operation. [ABSTRACT FROM AUTHOR]

Breast cancer is a leading cause of mortality and the most prevalent form of malignant tumor among women worldwide. Breast cancer cells exhibit an elevated glycolysis and altered glucose metabolism. Moreover, these cells display abnormal glycosylation patterns, influencing invasion, proliferation, metastasis, and drug resistance. Consequently, targeting glycolysis and mitigating abnormal glycosylation represent key therapeutic strategies for breast cancer. This review underscores the importance of protein glycosylation and glucose metabolism alterations in breast cancer. The current research efforts in developing effective interventions targeting glycolysis and glycosylation are further discussed. [ABSTRACT FROM AUTHOR]

Filter pruning has drawn much attention for releasing the limitations of person re-identification (ReID) in resource-constrained platforms due to the large parameters and high computational consumption. Filter pruning is depended on the given combination of pruning ratios to trim off unimportant filters, therefore, we define filter pruning as optimization problem. Existing strategies generally retrain all the pruning results and compare their performance iteratively, representing a major expenditure of time and effort. In this paper, we formally put forward a scheme of filter pruning based on evolutionary algorithms (EAFPruner) for ReID task for the first time, which automatically finds the optimal pruning structure in the pruning space with the better trade-off between computational cost and identification accuracy. Specifically, multiple evolutionary algorithms are adopted to solve the optimization problem by using mean average precision (mAP) as the fitness for filter-level sparsity. In addition, the adaptive batch normalization strategy is adopted to evaluate the pruned candidates generated by evolutionary algorithm without fine-tuning them to speed up training process. The experimental results are carried out on four person ReID datasets, and the number of parameters required by ResNet50 architecture is reduced by half while maintaining a comparable Rank-1 accuracy, which demonstrates the effectiveness of our proposed method. [ABSTRACT FROM AUTHOR]

Surface defect classification plays a very important role in industrial production and mechanical manufacturing. However, there are currently some challenges hindering its use. The first is the similarity of different defect samples makes classification a difficult task. Second, the lack of defect samples leads to poor accuracies when using deep learning methods. In this paper, we first design a novel backbone network, ResMSNet, which draws on the idea of multi-scale feature extraction for small discriminative regions in defect samples. Then, we introduce few-shot learning for defect classification and propose a Relation-Prototypical network (RPNet), which combines the characteristics of ProtoNet and RelationNet and provides classification by linking the prototypes distances and the nonlinear relation scores. Next, we consider a more realistic scenario where the base dataset for training the model and target defect dataset for applying the model are usually obtained from domains with large differences, called cross-domain few-shot learning. Hence, we further improve RPNet to KD-RPNet inspired by knowledge distillation methods. Through extensive comparative experiments and ablation experiments, we demonstrate that either our ResMSNet or RPNet proves its effectiveness and KD-RPNet outperforms other state-of-the-art approaches for few-shot defect classification. [ABSTRACT FROM AUTHOR]

Object detection is one of the most basic and important research tasks in the field of computer vision. The general trend in object detection has been to design large and over-parameterized models, which can achieve excellent performance. However, this comes at the expense of low speed, heavy computation and large amount of memory overhead, also makes object detection models more difficult to be applied on mobiles and embedded devices which have limited hardware resources and need real-time feedback. So there has been rising interest in building portable and efficient networks for object detection in the recent literature. The main contributions of this review include the following aspects. As far as we know, there are few reviews on efficient object detection CNNs. We systematically summarize the methods, models and evaluation metrics of efficient CNNs for object detection in recent years. We summarize and introduce some commonly used datasets for object detection. Finally, we point out some possible research directions and inspire some useful suggestions for the future work of efficient convolutional neural network. [ABSTRACT FROM AUTHOR]

The selective oxidative dehydrogenation of ethane (ODHE) is attracting increasing attention as a method for ethylene production. Typically, thermocatalysts operating at high temperatures are needed for C–H activation in ethane. In this study, we describe a low temperature (< 140 °C) photocatalytic route for ODHE, using O2 as the oxidant. A photocatalyst containing PdZn intermetallic nanoparticles supported on ZnO is prepared, affording an ethylene production rate of 46.4 mmol g–1 h–1 with 92.6% ethylene selectivity under 365 nm irradiation. When we employ a simulated shale gas feed, the photocatalytic ODHE system achieves nearly 20% ethane conversion while maintaining an ethylene selectivity of about 87%. The robust interface between the PdZn intermetallic nanoparticles and ZnO support plays a crucial role in ethane activation through a photo-assisted Mars-van Krevelen mechanism, followed by a rapid lattice oxygen replenishment to complete the reaction cycle. Our findings demonstrate that photocatalytic ODHE is a promising method for alkane-to-alkene conversions under mild conditions. The selective oxidative dehydrogenation of ethane is attracting increasing attention as a method for ethylene production. Here, PdZn supported on ZnO affords record-breaking photocatalytic ethane-to-ethylene conversion rate, emphasizing the pivotal role of the interface between PdZn and ZnO in the process. [ABSTRACT FROM AUTHOR]

The Actinomycetes Streptomyces lincolnensis is the producer of lincosamide-type antibiotic lincomycin, a widely utilized drug against Gram-positive bacteria and protozoans. In this work, through gene knockout, complementation, and overexpression experiments, we identified LcbR1 (SLINC_1595), a GntR family transcriptional regulator, as a repressor for lincomycin biosynthesis. Deletion of lcbR1 boosted lincomycin production by 3.8-fold, without obvious change in morphological development or cellular growth. The homologues of LcbR1 are widely distributed in Streptomyces. Heterologous expression of SCO1410 from Streptomyces coelicolor resulted in the reduction of lincomycin yield, implying that the function of LcbR1 is conserved across different species. Alignment among sequences upstream of lcbR1 and their homologues revealed a conserved 16-bp palindrome (-TTGAACGATCCTTCAA-), which was further proven to be the recognition motif of LcbR1 by electrophoretic mobility shift assays (EMSAs). Via this motif, LcbR1 suppressed the transcription of lcbR1 and SLINC_1596 sharing the same bi-directional promoter. SLINC_1596, one important target of LcbR1, exerted a positive effect on lincomycin production. As detected by quantitative real-time PCR (qRT-PCR) analyses, the expressions of all selected structural (lmbA, lmbC, lmbJ, lmbV, and lmbW), resistance (lmrA and lmrB) and regulatory genes (lmrC and lmbU) from lincomycin biosynthesis cluster were upregulated in deletion strain ΔlcbR1 at 48 h of fermentation, while the mRNA amounts of bldD, glnR, ramR, SLCG_Lrp, and SLCG_2919, previously characterized as the regulators on lincomycin production, were decreased in strain ΔlcbR1, although the regulatory effects of LcbR1 on the above differential expression genes seemed to be indirect. Besides, indicated by EMSAs, the expression of lcbR1 might be regulated by GlnR, SLCG_Lrp, and SLCG_2919, which shows the complexity of the regulatory network on lincomycin biosynthesis. Key points: • LcbR1 is a novel and conservative GntR family regulator regulating lincomycin production. • LcbR1 modulates the expressions of lcbR1 and SLINC_1596 through a palindromic motif. • GlnR, SLCG_Lrp, and SLCG_2919 can control the expression of lcbR1. [ABSTRACT FROM AUTHOR]

The relationship between Sn–Ag mineralization and mantle magmatism is a topic of high interest in current ore deposit research. Here, we investigate porphyry-, skarn-, and cassiterite-sulfide type Sn-polymetallic deposits associated with granitoids and vein-type Ag–Pb–Zn deposits hosted in sub-volcanic rocks in the southern Great Xing'an Range (SGXR), northeast China, as a case example. We use He, Ar, and S isotopes and isotopic end-member simulation calculations to determine the contribution of mantle-derived fluids/melts to the ore mineralization. Our He–Ar isotope data demonstrate that the ore-forming fluids are mixtures of shallow crust-derived fluid containing radiogenic 4He but no radiogenic 40Ar and magmatic fluids with mantle-derived 3He and 40Ar. The Pb–Zn–Ag deposits have a higher contribution of magmatic volatiles than the Sn-polymetallic deposits. Sulfide δ34S values of − 2.7 to − 0.6‰ in the Pb–Zn–Ag deposits are consistent with a magmatic sulfur source, whereas sulfides with δ34S values of − 12.2 to − 0.15‰ in the Sn-polymetallic deposits signal a possibly bimodal source of sulfur, i.e., crustal light sulfur mixed with magmatic sulfur. The noble gas compositions of the ore fluids are controlled by crustal thickness, high 3He fluxes (24 to 404 at/s/cm2), and low residence time (1 to 18 Myr) of He in the asthenosphere below the SGXR. Non-equilibrium open-system magma degassing is evidenced by the range of elevated values of 4He/40Ar* ratios (4.8–127). The 3He/heat ratio of the ore fluids from the Sn and Pb–Zn–Ag deposits overlap (0.01–0.76 × 10−2 cm3 STP J−1 (cubic centimeter at standard temperature and pressure per joule) and 0.02–1.08 × 10−2 cm3 STP J−1, respectively), indicating a consequence of conduction of mantle-derived heat across the magma-hydrothermal interface. Furthermore, an increasing abundance of Sn reserves in the SGXR deposits can be equated with an increase in the mantle-derived He component in the ore fluids. These findings suggest that a continuous flux of mantle-derived fluids/melts plays an essential role in Sn–Ag–Pb–Zn mineralization. [ABSTRACT FROM AUTHOR]

To perform object-specific tasks on the desktop, robots need to perceive different objects. The challenge is to calculate the pixel-wise mask for each object, even in the presence of occlusions and unseen objects. We take a step toward this problem by proposing a metric learning-based network called TRF-Net to perform desktop object instance segmentation. We design two ResNet-based branches to process the RGB and depth images separately. Then, we propose a Transformer-based fusion module called TranSE to fuse the features from both branches. This module also transfers the fused features to the decoder part, which helps generate fine-grained decoder features. After that, we propose a multi-scale feature embedding loss function called MFE loss to reduce the intra-class distance and increase the inter-class distance, which contributes to the feature clustering in embedding space. Due to the lack of large-scale real-world datasets for desktop objects, the proposed TRF-Net is trained with the synthetic dataset and tested with the small-scale real-world dataset. The target objects in the testing dataset do not present in the training dataset, ensuring the novelty of testing objects. We demonstrate that our method can produce accurate instance segmentation masks, outperforming other state-of-the-art methods on desktop object instance segmentation. [ABSTRACT FROM AUTHOR]

Engineering of crystal morphology affects the catalytic and adsorption properties of zeolitic materials. Considering the anisotropic diffusion of molecules derived from its topological features, MFI zeolite nanosheets with short b-axis thickness are highly desired materials to reduce diffusion resistance. However, the design and development of eco-friendly synthesis protocols with reasonable cost and high efficiency remain elusive. Herein, we reported a systematic study on the synthesis of MFI nanosheets using urea as an additive. Both silicalite-1 and ZSM-5 zeolites (MFI type framework structure) with controllable b-thicknesses ranging from 50–200 nm were achieved by optimizing the synthetic parameters including water content, urea and SDA concentrations. The concentration of hydroxide anions was found to dominate the crystallization kinetics compared with the counterpart tetrapropylammonium cations (TPA+). To facilitate the crystal growth of MFI zeolites in the presence of urea, the ratio OH−/SiO2 has to be higher than 0.2, independent of the TPA+ concentration. The role of urea in the assistance of plate-like crystal formation through the inhibition of (010) facet growth was revealed by electron microscopy and infrared (IR) spectroscopy analyses. The developed strategy for morphological engineering is not limited to the MFI-type zeolite and can be applied to other frameworks depending on the intrinsic properties of additive molecules and the interactions between them. [ABSTRACT FROM AUTHOR]

RGB-T tracking means that given the object position in the first frame, the tracker is trained to predict the position of the object in consecutive frames by taking full advantage of the complementary information of RGB and thermal infrared images. As the amount of data increases, unsupervised training has great potential for development in RGB-T tracking task. As we all know, features extracted from different convolutional layers can provide different levels information in the image. In this paper, we propose a framework for visual tracking based on the attention mechanism fusion of multi-modal and multi-level features. This fusion method can give full play to the advantages of multi-level and multi-modal information. Specificly, we use a feature fusion module to fuse these features from different levels and different modalities at the same time. We use cycle consistency based on a correlation filter to implement unsupervised training of the model to reduce the cost of annotated data. The proposed tracker is evaluated on two popular benchmark datasets, GTOT and RGB-T234. Experimental results show that our tracker performs favorably against other state-of-the-art unsupervised trackers with a real-time tracking speed. [ABSTRACT FROM AUTHOR]

This study investigated the trajectories in posttraumatic growth (PTG) among adolescents who survived from the Yancheng tornado in China, and explored the effects of posttraumatic stress symptoms (PTSS) on these trajectories. Participants (n = 246) finished 4 assessments at 6, 9, 12, and 18 months after the tornado. Growth mixture model and logistic regression were used to examine the heterogeneous trajectories and the role of PTSS for differentiating trajectories respectively. Two latent PTG trajectories were observed: group with decreasing PTG and group with fluctuant PTG, which might stem from the illusory component and the factual component of PTG respectively based on the two-component model; and adolescents with more PTSS had higher probabilities generating decreasing PTG, that is, illusory PTG. This study suggested differentiating PTG trajectories and related influencing factors to improve the post-disaster psychological interventions in a longitudinal perspective. [ABSTRACT FROM AUTHOR]

Incremental learning is one of the most important abilities of human beings. In the age of artificial intelligence, it is the key task to make neural network models as powerful as human beings, to achieve the ability to continuously acquire, fine-tune, and accumulate knowledge while simultaneously avoid catastrophic forgetting. In recent years, by virtue of deep neural networks, incremental learning has been attracting a great deal of attention in the field of computer vision. In this paper, we systematically review the current development of incremental learning and give the overall taxonomy of the incremental learning methods. Specifically, three kinds of mainstream methods, i.e., parameter regularization-based approaches, knowledge distillation-based approaches, and dynamic architecture-based approaches, are surveyed, summarized, and discussed in detail. Furthermore, we comprehensively analyze the performance of data-permuted incremental learning, class-incremental learning, and multi-modal incremental learning on widely used datasets, covering a broad of incremental learning scenarios for image classification and semantic segmentation. Lastly, we point out some possible research directions and inspiring suggestions for incremental learning in the field of computer vision. [ABSTRACT FROM AUTHOR]

The designed method aims to perform image classification tasks efficiently and accurately. Different from the traditional CNN-based image classification methods, which are greatly affected by the number of labels and the depth of the network. Although the deep network can improve the accuracy of the model, the training process is usually time-consuming and laborious. We explained how to use only a few of labels, design a more flexible network architecture and combine feature distillation method to improve model efficiency while ensuring high accuracy. Specifically, we integrate different network structures into independent individuals to make the use of network structures more flexible. Based on knowledge distillation, we extract the channel features and establish a feature distillation connection from the teacher network to the student network. By comparing the experimental results with other related popular methods on commonly used data sets, the effectiveness of the method is proved. The code can be found at https://github.com/ZhangXinba/Semi_FD. [ABSTRACT FROM AUTHOR]

To benefit tissue removal and postoperative rehabilitation, increased efficiency and accuracy and reduced operating force are strongly required in the osteotomy. A novel elliptical vibration cutting (EVC) has been introduced for bone cutting compared with conventional cutting (CC) in this paper. With the assistance of high-speed microscope imaging and the dynamometer, the material removals of cortical bone and their cutting forces from two cutting regimes were recorded and analysed comprehensively, which clearly demonstrated the chip morphology improvement and the average cutting force reduction in the EVC process. It also revealed that the elliptical vibration of the cutting tool could promote fracture propagation along the shear direction. These new findings will be of important theoretical and practical values to apply the innovative EVC process to the surgical procedures of the osteotomy. [ABSTRACT FROM AUTHOR]

Photothermal Fischer-Tropsch synthesis represents a promising strategy for converting carbon monoxide into value-added chemicals. High pressures (2-5 MPa) are typically required for efficient C-C coupling reactions and the production of C5+ liquid fuels. Herein, we report a ruthenium-cobalt single atom alloy (Ru1Co-SAA) catalyst derived from a layered-double-hydroxide nanosheet precursor. Under UV-Vis irradiation (1.80 W cm−2), Ru1Co-SAA heats to 200 °C and photo-hydrogenates CO to C5+ liquid fuels at ambient pressures (0.1-0.5 MPa). Single atom Ru sites dramatically enhance the dissociative adsorption of CO, whilst promoting C-C coupling reactions and suppressing over-hydrogenation of CHx* intermediates, resulting in a CO photo-hydrogenation turnover frequency of 0.114 s−1 with 75.8% C5+ selectivity. Owing to the local Ru-Co coordination, highly unsaturated intermediates are generated during C-C coupling reactions, thereby improving the probability of carbon chain growth into C5+ liquid fuels. The findings open new vistas towards C5+ liquid fuels under sunlight at mild pressures. Photothermal Fischer-Tropsch synthesis represents a promising strategy for converting CO into value-added chemicals. Here the authors report a Ru1Co single atom alloy catalyst for CO photo-hydrogenation to C5+ liquid fuels at ambient pressure, and confirm the role of Ru single atoms in promoting CO activation and C-C coupling whilst suppressing CHx* over-hydrogenation. [ABSTRACT FROM AUTHOR]

In this work, a series of iso-structural eulyite-type Ba3Yb(PO4)3:Tb3+ solid solution phosphors were synthesized via high-temperature solid-phase method. The corresponding unit cell structure, luminescence properties, thermal stability, and chromaticity coordinates of the as-obtained phosphors were systematically investigated. The crystal structure of Ba3Yb(PO4)3:Tb3+ was refined by GSAS program, which indicates the oxygen ions in Ba3Yb(PO4)3 are disordered at one C3 site, and the oxygen atoms are distributed at two sites. Under 378 nm ultraviolet excitation, the emission peak of Ba3Yb(PO4)3:Tb3+ appears in the range of 450–650 nm, and a characteristic green emission peak appears at 542 nm. Moreover, the luminous intensity of the Ba3Yb(PO4)3:14 at.% Tb3+ sample at 160 °C is 87.7% at room temperature, and the quantum yield is 55.8%, which reveals good thermal stability. Our results suggest that the Ba3Yb(PO4)3:Tb3+ phosphors are promising high-efficiency green light conversion materials for UV-LED applications. [ABSTRACT FROM AUTHOR]

The insufficient number of low-elevation observations is a limitation of the three-dimensional ionospheric computer tomography (CT) based on the global navigation satellite system (GNSS). To solve this problem, accurate prior information on the regional ionosphere must be obtained. However, it is difficult to explicitly and accurately express prior ionospheric information. This study uses compressed sensing (CS) for ionospheric tomography for the first time. Specifically, the electron density obtained from the international reference ionosphere is used to build a dictionary to fully integrate the prior information into the dictionary. Then, the electron density is reconstructed by using the compressive sampling matching pursuit method. Subsequently, the GNSS data of China (Region I) and Europe (Region II) were utilized to validate this proposed method, and the results are compared with ionosonde observations. The mean and standard deviation (SD) of the difference with respect to the ionosonde result are 41 and 22 km, respectively. The mean and SD of relative deviation were 16% and 9%, respectively. In Region II, the mean and SD of the deviation between the reversed peak electron density and the result of the ionosonde were 1.9 × 1010 m−3 and 8.1 × 1010 m−3, respectively. The mean and SD of the relative deviation were 3% and 13%, respectively. The mean and SD of the peak height deviation were 33 and 19 km, and the mean and SD of the relative deviation were 11% and 7%. The electron density distribution and variation in these two regions showed a local time dependence, and the horizontal gradient of the electron density in the latitude was greater than that in the longitude. Moreover, CT by CS is efficient, taking about 6 s per inversion based on an desktop computer with 16 GB RAM and Intel (R) Core (TM) i7-8700 CPU. [ABSTRACT FROM AUTHOR]

The current study aims to explore the long-time effect of resilience on academic self-concept and the mediating effect of different emotion regulation strategies in the relationship between resilience and academic self-concept. A total of 524 migrant children (fourth grade to sixth grade; 53.3% were boys) participated in the baseline study (T1). 419 participants completed the second assessment (T2, 3 months after T1), which were used for data analysis for this study. The results showed that resilience at T1 significantly and positively predicted academic self-concept at T2 (β = 0.201, p = 0.002). Reappraisal at T2 mediated the relationship between resilience at T1 and academic self-concept at T2 (β = 0.144, p < 0.001), whereas suppression at T2 did not (β = −0.011, p = 0.083). The study enriches the empirical research on academic self-concept of migrant children, and advises educational scholar and psychologist to improve students' academic self-concept by promoting their resilience and reappraisal ability. [ABSTRACT FROM AUTHOR]

Person re-identification (ReID) aims to identify the same person across multiple cameras. Gait recognition is the person ReID using human gait to identify a walking person, which is an effective identification technology with many advantages, such as remote identification and without invasion. State-of-the-art solutions solve the problem of extensive annotation skeleton information and labeling process by employing encoder-decoders to reconstruct skeleton data, which encodes gait feature with a fixed-length vector limiting the performance of this architecture. In this paper, we propose an end-to-end pipeline dubbed as SCA-GAN, which assembles spatial-channel attention with GAN-like framework to synthesize skeleton sequences reversely without labeled skeleton data. A disadvantage of traditional encoder-decoder architecture is that, because of the fixed-length latent vector encoded from the encoder, the decoder fails to learn the reasonable standard to generate imperfect samples. Therefore, we design a GAN-like framework for discriminative gait feature extraction via adversarial learning. In addition, for learning the rich global information of skeleton data, the information of skeleton is extracted via convolutional block embedding locality-aware attention mechanism. Specifically, a contrastive feature loss is constructed between the gait encoder and the gait decoder to minimize their pixel-wise distance explicitly. The proof-of-principle experiments and ablation study on several benchmarks prove that the proposed method significantly outperforms gait recognition counterparts in precision. [ABSTRACT FROM AUTHOR]

In order to realize high-accuracy heart rate (HR) estimation based on photoplethysmography (PPG) under the scenes of low signal-to-noise ratio (SNR) and co-frequency caused by motion artifacts (MAs), this paper presents a novel framework integrating two-stage variational mode decomposition (VMD) denoising method, noise compensation technology, and hidden Markov model (HMM)-based tracking algorithm. The two-stage VMD denoising method is designed to separate the HR signal from MA under low SNR scene. The noise compensation technology is applied to solve the problem of co-frequency. HMM-based HR tracking method is adopted to obtain the global optimization performance of HR estimation. The effectiveness and superiority of the proposed framework in solving problems of low SNR and co-frequency associated with motion artifacts have been verified by the HR estimation experiments carried out on three public high-SNR PPG databases (ISPC, BAMI I, BAMI II) and a self-built low-SNR database (WeData). Compared with the two classical frameworks namely joint sparse spectrum reconstruction (JOSS) and convolutional neural network-long short-term memory network (CNN-LSTM), the proposed framework obtains the lowest HR estimation errors (0.94 beats per minute (BPM) and 1.81 BPM respectively) on both BAMI 2 with the highest SNR (0.40 dB) and WeData with the lowest SNR (- 9.07 dB). For the low-SNR database Wedata, the average absolute error (AAE) decreases by more than 21 BPM. The research result of this study provides a solution for the realization of high-accuracy PPG-based HR estimation in exercise scenarios. [ABSTRACT FROM AUTHOR]

Iron tailings matrix is deficient in nutrients, and phytoremediation is one of the effective methods to improve tailings nutrients. The response of phytoremediation to tailings microorganisms remains to be studied. The present study analyzed rhizospheric soil of two kinds of plants bacterial diversity and community structure and their relationship with soil environmental factors. The results indicate that the rhizospheric soil bacteria species of Robinia pseudoacacia and Juniperus sabina were not significantly different from that of bare tailings, but rhizospheric soil bacterial community compositions and abundance were significantly different from that of bare tailings. Canonical correlation analysis (CCA) showed that soil alkali-hydrolyzable nitrogen (AN), soil total nitrogen (TN), and soil organic matter (SOM) were the main environmental factors affecting bacterial community diversity. Spearman's correlation analysis showed that AN, TN, and SOM were significantly positively correlated with the relative abundance of Gemmatimonadetes and Nitrospirae, and were significantly negatively correlated with that of Firmicutes, Fusobacteria, and Bacteroidetes. FAPROTAX function prediction showed that the functional microbial communities of rhizospheric soil of the two plants were significantly different from those of bare tailings. Overall, the findings support an increase of microbial diversity, SOM, and nitrogen in rhizospheric soil of revegetated tailings compared to bare tailings. These results provide theoretical support for the development and application of phytoremediation in abandoned mines. [ABSTRACT FROM AUTHOR]

The ability to noninvasively detect and monitor the growth of orthotopic liver transplantation tumors is critical for replicating advanced colorectal cancer liver metastases (CRLMs) in animal models. We assessed the use of high-resolution ultrasound (HRU) to monitor CRLMs transplanted using various cell concentrations. Sixty BALB/c female mice were randomly divided into 3 groups, and murine colonic CT26 cells were injected into the left liver lobe at concentrations of 1 × 102 (group 1), 1 × 103 (group 2), or 1 × 104 (group 3). Tumor presentation, location, number, size, shape, and echogenicity were assessed daily with 24-MHz center frequency HRU starting 6 days after injection. Animals were sacrificed when the largest tumor was ≥ 1 cm in diameter. Sensitivity, specificity, and area under curve (AUC) of CRLMs diagnosed with HRU were calculated using receiver operating characteristic curve analysis. In group 1, 94% of mice formed < 5 tumors, and 41% formed a single tumor. Tumors were first detected with HRU on day 12 in group 1, day 10 in group 2, and day 7 in group 3; tumor volume doubling times were 14–15 days, 11–12 days, and 7–8 days, respectively. With a long diameter threshold of 2.4 mm, diagnostic sensitivity and specificity of HRU were 94.1% and 88.7%, respectively, and the AUC was 0.962. These findings suggest that HRU can be used to accurately detect and monitor the growth of CRLMs in an orthotopic transplantation mouse model, especially when a lower concentration of cells is used. [ABSTRACT FROM AUTHOR]

Correlations in facial action units (AUs) convey significant information for AU detection yet have not been thoroughly exploited. Most existing methods learn the regional correlation distribution of each AU, or reason the dependencies among AUs. However, these methods typically either predefine the correlations based on prior knowledge, which often ignores useful information, or directly learn the correlations guided by AU detection, which often includes irrelevant information. To resolve these limitations, we propose a novel hybrid relational reasoning framework for AU detection. In particular, we propose to adaptively reason pixel-level correlations of each AU, under the constraint of predefined regional correlations by facial landmarks, as well as the supervision of AU detection. Moreover, we propose to adaptively reason AU-level correlations using a graph convolutional network, by considering both predefined AU relationships and learnable relationship weights. Our framework is beneficial for integrating the advantages of correlation predefinition and correlation learning. Extensive experiments demonstrate that our approach (i) soundly outperforms the state-of-the-art AU detection methods on the challenging BP4D, DISFA, and GFT benchmarks, and (ii) can precisely reason the regional correlation distribution of each AU. [ABSTRACT FROM AUTHOR]

Person re-identification (Re-ID), aiming to retrieve interested people through multiple non-overlapping cameras, has caused concerns in pattern recognition communities and computer vision in recent years. With the continuous promotion of deep learning, the research on person Re-ID is more and more extensive. In this paper, we conduct a comprehensive review of the advanced methods and divide them into three categories from coarse to fine: (1) global-based methods, which are based on whole images to obtain discriminative features; (2) part-based methods, which focus on image regions to extract detailed information; (3) multiple granularities-based methods, which combine advantages of the above two categories. For each category, we further classify it according to popular research tools. Then, we give the evaluation of some typical models on a set of benchmark datasets and compare them in detail. We also introduce some widely used training tricks. The methods mentioned in this paper were published in 2011-2021. By discussing their advantages and limitations, we provide a reference for future works. [ABSTRACT FROM AUTHOR]

Video person re-identification (video Re-ID) is a key technology applied to video surveillance and security. Typical person re-identification is designed to retrieve the correct match of the target image (query) from gallery images, while video Re-ID extends this to query from gallery videos. The main factors affecting the video Re-ID model are: (i) a high-quality frame-level feature extractor, and (ii) temporal modeling that combines frame-level features into a feature for retrieval. In this work, we use ShuffleNet V2-based lightweight algorithm for video Re-ID, which can meet the demand for practical application and solve the problem of high consumption for computing resources, and maintain high performance. At the same time, the lightweight space attention mechanism Spatial Group-wise Enhance (SGE) module is used to view the person in more detail, which makes the feature representation more compact and effectively improves the retrieval accuracy. Finally, we design an Online Difference Discrimination (ODD) module to measure the feature gap between video frames, and use this module to make different temporal modeling for different quality video sequences. Experiments on three datasets (i.e., iLIDS-VID, PRID2011 and MARS) show that our method is competitive with state-of-the-art methods. [ABSTRACT FROM AUTHOR]

The Hengduan Mountains Region (HMR) is one of the areas that experience the most frequent geological hazards in China. However, few reports are available that address the geological hazard susceptibility of the region. This study developed six machine learning models to assess the geological hazard susceptibility. The results show that areas with medium and high susceptibility to geological hazards as a whole account for almost 21% of the total area, while both are 18% when it comes to the single hazard of landslide and rockfall respectively. Medium and high geological hazard susceptibility is found in three parts of the HMR with different characteristics: (1) the central and southern parts, where the population of the region concentrates; (2) the northern part, where higher geological hazard susceptibility is found along the mountain ranges; and (3) the junction of Tibet, Yunnan, and Sichuan in the eastern part, which is prone to larger-scale geological hazards. Of all the potential influencing factors, topographic features and climatic variables act as the major driving factors behind geological hazards and elevation, slope, and precipitation are crucial indicators for geological hazard susceptibility assessment. This study developed the geological hazard susceptibility maps of the HMR and provided information for the multi-hazard risk assessment and management of the region. [ABSTRACT FROM AUTHOR]

Due to the different shooting angles, altitudes and scenes, remote sensing images contain many complex backgrounds and multi-scale objects. Moreover, objects in remote sensing images are much smaller relative to the backgrounds, easily occluded by buildings and trees. These cause difficult feature extraction and increase the intra-class diversity of objects, making object detection on remote sensing images more challenging. In this paper, we propose a novel remote sensing image object detection method (SHDet) based on spatial hierarchy perception component (SHPC) and hard samples metric learning (HSML). We design a SHPC to extract the feature under the different spatial hierarchies and learn the contribution weights between feature channels to enhance the feature representation. HSML is proposed to narrow the feature differences of hard samples in the same category, reducing the error detection caused by intra-class diversity. Besides, we decouple the complex background to build the pre-training datasets for pre-training the object detection model, strengthening the object feature learning. The experiments carried out on two widely used remote sensing datasets (NWPU VHR-10 and DOTA-v1.5) show that the proposed method has better detection performance compared with several state-of-the-art object detection methods. [ABSTRACT FROM AUTHOR]

A series of single-phase white-emitting phosphors Ba3Bi(PO4)3 (BBP):8 at.% Dy, x at.% Eu were successfully synthesized through high temperature solid-phase method. The luminescence properties of BBP:8 at.% Dy, x at.% Eu were systematically investigated. Experimental results indicate the doping of Eu3+ ions can effectively compensate the red light component in the emission spectrum and tune the correlated color temperature of BBP:8 at.% Dy, x at.% Eu from cool white light (5097 K) to warm white light (3661 K). Moreover, BBP:8 at.% Dy3+, 0.8 at.% Eu3+ phosphor exhibits excellent thermal stability and the fluorescence intensity at 150 °C can still maintain 76% of that at 25 °C, while the activation energy is determined to be ~ 0.23 eV. In addition, the quantum efficiency of BBP:8 at.% Dy3+, 0.8 at.% Eu3+ phosphor were measured to be 61%. These results suggest that BBP:8 at.% Dy3+, 0.8 at.% Eu3+ can be a NUV-pumped single-phase white phosphors with excellent comprehensive performances. [ABSTRACT FROM AUTHOR]

With the increasing thermal challenges of electronic equipment, topology optimization of heat sinks has aroused great research interest in recent years. However, most related researches deal with heat transfer problems of the laminar flow. Only several papers deal with heat transfer problems of the turbulent flow which exhibits better heat dissipation capacity. This paper intends to demonstrate the topology optimization of turbulent forced convective heat sinks with a computational-cheap multi-layer thermofluid model. Based on the assumption of a hydrodynamically fully developed turbulent flow, a single-layer flow model is derived from the 3D turbulent flow model. Viscous effects in the out-of-plane direction are modeled as the additional terms in the derived momentum and transportation equations. Similarly, based on the assumption of adaptive temperature profiles in the out-of-plane direction, heat fluxes between adjacent layers are modeled and a multi-layer heat transfer model is constructed. Further, the porosity field is introduced to describe the channel's topology and the corresponding topology optimization scheme is presented. Topology optimizations of a heat sink for coil-cooling are carried out to reduce the average temperature rise of the heat source under a maximum volume fraction constraint. The accuracy of the multi-layer model is evaluated via 3D thermofluid simulation of optimized and reference designs. Topology-optimized heat sinks exhibit better thermal performance and provide up to a 32.39% reduction of the average temperature rise, compared with the reference designs. Additionally, a comparative study is also provided between designs optimized in heat transfer problems under laminar and turbulent flows. [ABSTRACT FROM AUTHOR]

Photothermal CO2 reduction is an efficient and sustainable catalytic path for CO2 treatment. Here, we successfully fabricated a novel series of Ni-based catalysts (Ni-x) via H2 reduction of NiAl-layered double hydroxide nanosheets at temperatures (x) ranging from 300 to 600 °C. With the increase of the reduction temperature, the methane generation rate of the Ni-x catalyst for photothermal CO2 hydrogenation gradually increased under ultraviolet-visible-infrared (UV-vis-IR) irradiation in a flow-type system. The Ni-600 catalyst showed a CO2 conversion of 78.4%, offering a CH4 production rate of 278.8 mmol·g−1h−1, with near 100% selectivity and 100 h long-term stability. Detailed characterization analyses showed metallic Ni nanoparticles supported on amorphous alumina are the catalytically active phase for CO2 methanation. This study provides a possibility for large-scale conversion and utilization of CO2 from a sustainable perspective. [ABSTRACT FROM AUTHOR]

Person re-identification (ReID) is mainly aimed at establishing correct identity correspondence among moving person collected by multiple cameras. Extending labeled data sets with pseudo-labels is one of the common methods of ReID. However, single evaluation standards and fixed screening pseudo-label methods make pseudo-labels gradually weaken their update rate. Based on that, we propose a semi-supervised adaptive stepwise learning (SSAS) method for accelerating the update of pseudo-labels. Using the concept of Kullback–Leibler divergence, a more global pseudo-label update idea (GPLU) is proposed, an evaluation criterion of pseudo-labels is designed to satisfy two conditions: The first is to use simple tracklets as pseudo-label data in the early stage, and the second is to gradually add complex and diverse tracklets as pseudo-label data in the iterative process. Our proposed adaptive pseudo-label screening strategy steadily improves the recognition accuracy of ReID. In addition, we conduct extensive experiments on canonical data sets and the evaluation results suggest the superiority of our method. [ABSTRACT FROM AUTHOR]

Deep learning methods show strong ability in extracting high-level features for images in the field of person re-identification. The produced features help inherently distinguish pedestrian identities in images. However, on deep learning models over-fitting and discriminative ability of the learnt features are still challenges for person re-identification. To alleviate model over-fitting and further enhance the discriminative ability of the learnt features, we propose siamese pedestrian alignment networks (SPAN) for person re-identification. SPAN employs two streams of PAN (pedestrian alignment networks) to increase the size of network inputs over limited training samples and effectively alleviate network over-fitting in learning. In addition, a verification loss is constructed between the two PANs to adjust the relative distance of two input pedestrians of the same or different identities in the learned feature space. Experimental verification is conducted on six large person re-identification data sets and the experimental results demonstrate the effectiveness of the proposed SPAN for person re-identification. [ABSTRACT FROM AUTHOR]

Oxygen deficiencies are well recognized to have a detrimental effect on the performance of spinel LiMn2O4-δ as the cathode for non-aqueous systems, whereas the influences have not been explored for aqueous systems yet. Here, we successfully introduce oxygen deficiencies into LiMn2O4-δ by calcination of oxygen deficiency-free LiMn2O4 at 850 °C for 2 h followed by quenching in air. LiMn2O4-δ primary particles show the octahedral shape, the sizes of about 400 nm, and the larger lattice constant of 8.240 Å, in contrast with the irregular shapes, the sizes of about 100 nm, and the lattice constant of 8.231 Å. In the non-aqueous system, the performances for LiMn2O4-δ are inferior compared to LiMn2O4, as reported previously. However, both LiMn2O4-δ||polyimide and LiMn2O4||polyimide aqueous full cells exhibit the discharge capacity of about 105 mAh g−1 at 100 mA g−1 and the capacity retentions of 87% and 60% after 100 cycles. The plateau of H2O decomposition observed at about 1.85 V for LiMn2O4 leads to much poorer performance which is absent for LiMn2O4-δ. This study demonstrates different requirements by the cathodes in the non-aqueous and aqueous systems and offers a new idea for improving the performance and stability of the electrodes for aqueous cells. [ABSTRACT FROM AUTHOR]

With the increasing development of topology optimization theory in thermal-fluid problems, it has been widely applied in design problems of cooling structures. However, little researches have concurrently considered the design-dependent mechanical problems of optimized structures. This paper deals with porosity-based topology optimization of forced convection planar heat sinks considering out-of-plane design-dependent deformation problems. The optimization is performed on the middle cross-section and the mechanical behaviors of the upper cover plate are considered. Governing equations of fluid flow, heat transfer, and structural mechanic fields are constructed. Particularly, to concurrently model design-dependent loads and boundary conditions, a spring term with design-dependent stiffnesses and a design-dependent pressure term are introduced into the out-of-plane mechanical model. Numerical examples in fluid-mechanical decoupled and weakly-coupled cases are given to demonstrate the validity and potential of the present approach. Additionally, the presented scheme is also evaluated in cases considering the global mechanical behavior and cases considering the local mechanical behavior. Comparisons and verifications show that the proposed approach is effective in yielding planar heat sinks with optimized heat transfer capabilities and well-controlled out-of-plane mechanical behaviors. [ABSTRACT FROM AUTHOR]

This paper investigates the topology optimization of planar cooling channels using a low-cost multilayer thermofluid model. A novel three-layer model including the upper/lower cover-plate layers and the central solid-fluid mixing layer is proposed. The flow boundary layer effect and the heat transfer effect in the thickness direction are modeled as the flow coupling and thermal coupling effects between adjacent layers, respectively. Particularly, in order to estimate more accurate temperature fields, the constructed three-layer heat transfer model in the solid-fluid mixing channel is derived based on the assumption of adaptive temperature profiles in the thickness direction. Further, based on the three-layer thermofluid model, the porosity field is introduced to describe the channel's topology, and the corresponding topology optimization scheme is presented. Several optimized channels under different constraints and boundary conditions are shown and discussed in comparison. Optimized channels show streamlined boundaries and reasonable layouts and exhibit competitive heat dissipation performance. Parametric studies are conducted to analyze the effectiveness and reliability of the topology optimization scheme. Further, to evaluate the accuracy and efficiency of the proposed three-layer model, optimized channels are simulated with the full 3D model and previous low-cost models. The proposed three-layer model exhibits good consistency with the full 3D model while achieves great improvement in efficiency. [ABSTRACT FROM AUTHOR]

Despite considerable studies focused on the symptoms of posttraumatic stress disorder (PTSD), little is understood about how symptoms of PTSD naturalistically change over time. Using network analyses approaches, the current study aimed to understand the nature of the association between PTSD symptoms at different time points among adolescents who experienced an earthquake. This study enrolled 900 youth survivors who completed 3 assessments with the Child PTSD Symptom Scale at 1 year, 1.5 years, and 2 years after the Wenchuan earthquake. A graphical Gaussian model (GGM) was used to investigate how symptom networks changed across these time points and to identify the symptoms that were the most central within the network. Results from GGM indicated that different symptoms were observed to have highest centrality at different time points. Feeling distant or cut off from others, avoid thoughts and feelings about the trauma, and feeling irritable or having angry outbursts appeared as the node with highest centrality at 1 year (T1), 1.5 years (T2), and 2 years (T3) post-earthquake, respectively. [ABSTRACT FROM AUTHOR]

These compounds are common in several organisms; however, this is the first report that describes them as components of I H. bachu. i The antioxidant activity [[14]] of HBE was detected by ABTS (Fig. In this study, the chemical components of the ethanol extract of I H. bachu i (HBE) were analyzed, and their antioxidant and anticancer activities were evaluated I in vitro i . Among all the cells, the inhibitory effect on HCT116 cells was the most pronounced, reaching 69% with 0.8 mg/mL HBE, whereas the inhibition rate on BGC was the lowest (40%), and on other cancer cells (HepG2 and RAW264.7) it was nearly 50%. Eleven compounds were obtained, and all these compounds were first isolated from I Helvella bachu i . [Extracted from the article]

With the increasing dissipated power levels of electronic equipment, heat sink design problems are significant aspects of some industrial design fields. Topology optimization theory has attracted much research interest in thermal-fluid problems to design high-performance heat sinks. These researches focus on optimizing the internal flow channel of heat sinks under predefined inlet and outlet schemes. However, inlet/outlet schemes also exhibit great influence on the performance of optimized heat sinks. This paper intends to demonstrate a concurrent optimization method of the internal flow channel, inlets, and outlets in forced convection heat sinks. The porosity field and moving morphable components (MMCs) are adopted to describe structural topologies of the internal channel design domain and the inlet/outlet design domains, respectively. The internal porosity field and the shape/position parameters of MMCs are applied as design variables. A conventional inlet/outlet component is presented and the corresponding topology description transformation function is constructed to convert the inlet/outlet topology into corresponding porosity fields. Further, the thermal-fluid coupling problems are solved through porosity-interpolated fluid flow and heat transfer governing equations. To ensure accurate inlet flow and temperature conditions during the topology iteration, continuous design-dependent inlet governing equations are established under inlet pressure, velocity, and flow flux boundary conditions. Design variables are concurrently optimized based on the sensitivity information. The proposed method is shown in detail by taking design problems of water-cooled heat sinks as examples. Various numerical examples are presented to validate the applicability and efficiency of this method. [ABSTRACT FROM AUTHOR]

In recent years, convolutional neural networks (CNN) have been extensively used for generic object detection due to their powerful feature extraction capabilities. This has hence motivated researchers to adopt this technology in the field of remote sensing. However, remote sensing images can contain large amounts of noise, have complex backgrounds, include small dense objects as well as being susceptible to weather and light intensity variations. Moreover, from different shooting angles, objects can either have different shapes or be obscured by structures such as buildings and trees. Due to these, effective features extraction for proper representation is still very challenging from remote sensing images. This paper therefore proposes a novel remote sensing image object detection approach applying a fusion-based feature reinforcement component (FB-FRC) to improve the discrimination between object feature. Specifically, two fusion strategies are proposed: (i) a hard fusion strategy through artificially-set rules, and (ii) a soft fusion strategy by learning the fusion parameters. Experiments carried out on four widely used remote sensing datasets (NWPU VHR-10, VisDrone2018, DOTA and RSOD) have shown promising results where the proposed approach manages to outperform several state-of-the-art methods. [ABSTRACT FROM AUTHOR]

Image captioning is a task generating the natural semantic description of the given image, which plays an essential role for machines to understand the content of the image. Remote sensing image captioning is a part of the field. Most of the current remote sensing image captioning models failed to fully utilize the semantic information in images and suffered the overfitting problem induced by the small size of the dataset. To this end, we propose a new model using the Transformer to decode the image features to target sentences. For making the Transformer more adaptive to the remote sensing image captioning task, we additionally employ dropout layers, residual connections, and adaptive feature fusion in the Transformer. Reinforcement Learning is then applied to enhance the quality of the generated sentences. We demonstrate the validity of our proposed model on three remote sensing image captioning datasets. Our model obtains all seven higher scores on the Sydney Dataset and Remote Sensing Image Caption Dataset (RSICD), four higher scores on UCM dataset, which indicates that the proposed methods perform better than the previous state of the art models in remote sensing image caption generation. [ABSTRACT FROM AUTHOR]

Heart failure (HF) is a worldwide disease with high levels of morbidity and mortality. The pathogenesis of HF is complicated and involves imbalances in hormone and electrolyte. B-type natriuretic peptide (BNP) has served as a biomarker of HF severity, and in recent years, it has been used to treat the disease, thanks to its cardio-protective effects, such as diuresis, natriuresis, and vasodilatation. In stage C/D HF, symptoms are severe despite elevated BNP. Disturbances in Ca2+ homeostasis are often a dominating feature of the disease, causing Ca2+-regulatory protein dysfunction, including reduced expression and activity of sarcoplasmic reticulum Ca2+-ATPase2a (SERCA2a), impaired ryanodine receptors (RYRs) function, intensive Na+-Ca2+ exchanger (NCX), and downregulation of S100A1. The relationship between natriuretic peptides (NPs) and Ca2+-regulatory proteins has been widely studied and represents important mechanisms in the etiology of HF. In this review, we present evidence that BNP may regulate Ca2+-regulatory proteins, in particular, suppressing SERCA2a and S100A1 expression. However, relationships between BNP and other Ca2+-regulatory proteins remain vague. [ABSTRACT FROM AUTHOR]

Image semantic segmentation is one of the most important tasks in the field of computer vision, and it has made great progress in many applications. Many fully supervised deep learning models are designed to implement complex semantic segmentation tasks and the experimental results are remarkable. However, the acquisition of pixel-level labels in fully supervised learning is time consuming and laborious, semi-supervised and weakly supervised learning is gradually replacing fully supervised learning, thus achieving good results at a lower cost. Based on the commonly used models such as convolutional neural networks, fully convolutional networks, generative adversarial networks, this paper focuses on the core methods and reviews the semi- and weakly supervised semantic segmentation models in recent years. In the following chapters, existing evaluations and data sets are summarized in details and the experimental results are analyzed according to the data set. The last part of the paper is an objective summary. In addition, it points out the possible direction of research and inspiring suggestions for future work. [ABSTRACT FROM AUTHOR]

Photodynamic therapy (PDT) is a promising non-invasive therapy approach for various diseases including malignant tumor. The process of PDT involves three interrelated aspects, namely photosensitizer (PS), light source, and oxygen, among which PS is the decisive factor that determines its anticancer efficiency. There exist some defects in currently applied PDT, such as inadequate production of reactive oxygen species (ROS), poor penetration of exciting light, insufficient oxygen supply, and nonselective distribution of PS. With unique physicochemical and optical properties, two-dimensional nanomaterials (2DNMs) have aroused great interest in biomedical fields. 2DNMs-based PDT is promising to significantly improve antitumor efficacy compared to conventional PDT. In this review, we will firstly introduce the underlying mechanism of PDT and how 2DNMs are absorbed and distribute inside tumor cells. After that, we will not only illustrate how 2DNMs-based PDT can enhance tumor-killing efficacy and minimize side-effects through conquering the above-mentioned defects of conventional PDT and the preparation process of 2DNMs, but also elaborate recent advances about 2DNMs-based PDT. Lastly, we will summarize the challenges and future prospects of 2DNMs-based PDT. [ABSTRACT FROM AUTHOR]

Benefited from the rapid development of additive manufacturing (AM), inhomogeneous cellular structures have attracted many interests for their superior structural and functional performance. Recently proposed density-based design methods have been shown to provide great computational efficiency and obtain structures with excellent performance. To achieve better structural performance while considering AM constraints, an improved density-based design method which introduces solid and void units into the design domain is proposed in this paper. First, based on homogenization theory and solid-body analysis, unit parameters of different preset unit relative densities are determined. And a unit effective property interpolation model is constructed. Then, the macro relative density layout is optimized with density methods. In the optimization process, an efficient density filter is proposed to increase the optimization domain and satisfy minimal feature size constraint. Finally, the structure reconstruction algorithm automatically constructs the optimized cellular structure based on the unit and density information obtained in the first two processes. Numerical examples show that the proposed method efficiently obtains inhomogeneous cellular structures with better performance, compared with existing density-based methods. [ABSTRACT FROM AUTHOR]

The aim of this study was to investigate the effects of dietary konjac oligosaccharide (KOS) supplements on the serum immune parameters, intestine microbiota, and intestine immunity of Schizothorax prenanti. Fish were fed a control diet (basal diet without KOS) or a diet with KOS supplements (0.4% KOS g kg−1 [low-KOS] and 1.6% KOS g kg−1 [high-KOS]. Fish on the high-KOS diet had significant increases in serum alternative hemolytic complement activity, complement 3 content, and nitric oxide content and a significant decrease in malondialdehyde content. Total superoxide dismutase activity was significantly in fish on KOS-supplemented diets, regardless of inclusion level. The lactic acid bacteria levels were significantly increased in fish on the high-KOS diet, while the Aeromonas populations were significantly decreased in fish on the KOS-supplemented diets, regardless of inclusion level. Lysozyme activity in both the mid and distal intestine was significantly enhanced by the high-KOS diet. Only the high-KOS diet significantly increased the acid phosphatase activity in the distal intestine. Regarding immune-related gene expressions, the high-KOS diet was observed to significantly induce the expression of Toll-like receptor 22 in the both anterior and distal intestine. Tumor necrosis factor-alpha mRNA level in the whole intestine was significantly up-regulated in fish on the KOS-supplemented diets, regardless of the inclusion level. The expression of Interleukin 1beta in both mid and distal intestine was significantly promoted by the KOS-supplemented diets, regardless of inclusion level. [ABSTRACT FROM AUTHOR]