Search

Deep learning models have been widely applied in the field of long-term forecasting has achieved significant success, with the incorporation of inductive bias such as periodicity to model multi-granularity representations of time series being a commonly employed design approach in forecasting methods. However, existing methods still face challenges related to information redundancy during the extraction of inductive bias and the learning process for multi-granularity features. The presence of redundant information can impede the acquisition of a comprehensive temporal representation by the model, thereby adversely impacting its predictive performance. To address the aforementioned issues, we propose a De-Redundant Multi-Period Hybrid Modeling Network (DPHM-Net) that effectively eliminates redundant information from the series inductive bias extraction mechanism and the multi-granularity series features in the time series representation learning. In DPHM-Net, we propose an efficient time series representation learning process based on a period inductive bias and introduce the concept of de-redundancy among multiple time series into the representation learning process for single time series. Additionally, we design a specialized gated unit to dynamically balance the elimination weights between series features and redundant semantic information. The advanced performance and high efficiency of our method in long-term forecasting tasks against previous state-of-the-art are demonstrated through extensive experiments on real-world datasets. [ABSTRACT FROM AUTHOR]

Abnormal electricity usage detection is the process of discovering and diagnosing abnormal electricity usage behavior by monitoring and analyzing the electricity usage in the power system. How to improve the accuracy of anomaly detection is a popular research topic. Most studies use neural networks for anomaly detection, but ignore the effect of missing electricity data on anomaly detection performance. Missing value completion is an important method to improve the quality of electricity data and to optimize the anomaly detection performance. Moreover, most studies have ignored the potential correlation relationship between spatial features by modeling the temporal features of electricity data. Therefore, this paper proposes an electricity anomaly detection model based on multi-feature fusion and contrastive learning. The model integrates the temporal and spatial features to jointly accomplish electricity anomaly detection. In terms of temporal feature representation learning, an improved bi-directional LSTM is designed to achieve the missing value completion of electricity data, and combined with CNN to capture the electricity consumption behavior patterns in the temporal data. In terms of spatial feature representation learning, GCN and Transformer are used to fully explore the complex correlation relationships among data. In addition, in order to improve the performance of anomaly detection, this paper also designs a gated fusion module and combines the idea of contrastive learning to strengthen the representation ability of electricity data. Finally, we demonstrate through experiments that the method proposed in this paper can effectively improve the performance of electricity behavior anomaly detection. • Integrating temporal and spatial features to jointly accomplish electricity anomaly detection. • Designing a gated fusion module to capture the dependencies between multi-features. • Introducing contrastive learning to enhance the representation ability of electricity data. [ABSTRACT FROM AUTHOR]

Knowledge graph (KG) has been widely used in the field of recommender systems. There are some nodes in KG that guide the occurrence of interaction behaviors. We call them guided nodes. However, the current application doesn't take into account the guided nodes in KG. We explore the utility of guided nodes in KG. It is applied in repository recommendations. In this paper, we propose an end-to-end framework, namely Guided Node Graph Convolutional Network (GNGCN), which effectively captures the connections between entities by mining the influence of related nodes. We extract samples of each entity in KG as their guided nodes and then combine the information and bias of the guided nodes when computing the representation of a given entity. The guided nodes can be extended to multiple hops. We evaluate our model on a real-world Github dataset named Github-SKG and music recommendation dataset, and the experimental results show that the method outperforms the recommendation baselines and our model is much lighter than others. [ABSTRACT FROM AUTHOR]

Knowledge graph-aware recommendation has become an important research topic in recent years. The user preference representation, which preserves the user's taste towards items (e.g., movies, books.), is obtained through aggregating the information of entities or attributes in knowledge graphs directly. However, the fine-grained heterogeneity information, which can be derived from the groups of items or entities, remains barely exploited in the process of encoding the user interaction intention for the items. To fill up this gap, we propose a MultistageClustering-basedHierarchicalAttention (McHa) model to capture the user preference representation. In our work, we first group the items and their neighboring entities in the knowledge graph into item clusters and entity clusters (jointly referred to as multistage clusters), respectively. Then, the user preference representation is obtained by hierarchically aggregating the heterogeneity information derived from the multistage clusters with the weights generated by the hierarchical attention layers. We conduct extensive experimental comparisons with baselines and the variants. The experimental results indicate that McHa has achieved state-of-the-art performance on three benchmark datasets in two scenarios. [ABSTRACT FROM AUTHOR]

How to accurately predict the future medical treatment behaviors of patients from the historical health insurance data has become an important research issue in healthcare. In this paper, an Attention-based Bidirectional Gated Recurrent Unit (AB-GRU) medical treatment migration prediction model is proposed to predict which hospital patients will go to in the future. The model considers the impact of medical visit on the future medical behavior, on the basis of Bidirectional Gated Recurrent Unit (B-GRU) framework, we introduce an attention mechanism to determine the strength of hidden state at different moments, which can improve the predictive performance of the model. Due to medical treatment in different places has an important impact on the distribution of health insurance funds, the individual patient would be expected to the appropriate hospital and get the appropriate medical treatment. Therefore, when medical treatment prediction has been completed, this paper proposes a Similarity and Double-layer CNN-based (SD_CNN) medical treatment migration recommendation model. The model introduces a CNN framework to achieve patient similarity learning, and compares similarities to recommend whether patients need medical treatment migration. Finally, the experiment demonstrates that the model proposed in this paper is more accurate than other models. [ABSTRACT FROM AUTHOR]

[Display omitted] • Support has a great impact on plasma catalytic oxidation of propane. • Strong metal-support interactions govern the synergistic performance of Pt/CeMnO x. • Plasma activates oxygen whilst exciting the oxidation ability of Pt/CeMnO x. • Intermediates can be rapidly oxidized on surface of Pt/CeMnO x assisted with plasma. • Synergistic mechanism of propane decomposition by Pt/CeMnO x is proposed. The elimination of light alkanes has been the subject of extensive research. Rational tuning of the surface electronic structure of catalysts can optimize the adsorption and activation behavior of reactant molecules. Therefore, the role of metal-support interactions in plasma catalysis deserves in-depth investigation. In this study, a simple strategy was employed to synthesize Pt-based catalysts loaded on Ce-based binary oxide supports. The interactions between Pt noble metal and supports were regulated by varying the metal element type of the supports and the ratio of the support metal elements. Among them, 0.3 %Pt/CeMn 0.2 O x catalyst was the most effective for the synergistic plasma degradation of propane. Various characterizations showed that the redox ability and Pt species dispersion of 0.3 %Pt/CeMn 0.2 O x catalyst were significantly enhanced by strong metal-support interactions. It induced more Pt0 and Ce3+ on the surface for the activation of oxygen or water vapor to generate surface reactive oxygen species. Moreover, the plasma in turn facilitated the electron transfer of Pt/CeMn 0.2 O x catalyst and enhanced the metal-support interactions. The reactive oxygen species generated by plasma participated in the oxidation of propane and the oxygen cycle of Pt/CeMn 0.2 O x catalyst, thus enhancing the synergistic catalysis. The present work provides insight into the design of catalysts with high plasma catalytic synergy. [ABSTRACT FROM AUTHOR]

Leveraging knowledge graphs (KGs) to enhance recommender systems has gained considerable attention, with researchers obtaining user preferences by aggregating entity pairs with explicit relations in KGs via graph convolutional networks (GCNs). Existing approaches currently overlook many entity pairs without relations, which, however, may have potentially useful information. To address this issue, we propose a novel relation-aware attentional GCN (RAAGCN) with the following improvements over vanilla GCNs: (1) it aggregates all entity pairs with and without explicit relations and (2) it distinguishes the importance of different relational context information. Based on the proposed RAAGCN, we further propose a user preference and item attractiveness capturing model (UPIACM) for KG-based recommendation. In the UPIACM, the user preference is decomposed into interest and rating preferences. The interest preference is the user's interest taste toward the items with specific features, while the rating preference reflects the intention of rating high or low. Additionally, our model accounts for item attractiveness, which reflects an item's popularity among users. Additionally, we incorporate a gated filtering mechanism to further improve our model's performance. Through extensive experiments, we show that the proposed UPIACM outperforms state-of-the-art baseline methods. • Designing relation-aware attentional GCN to aggregate more fine-gained information. • Proposing KG-based UPIACM to capture user preference and item attractiveness. • Integrating gated filtering mechanism to filter out the possible impurity information. • Extensive experiments with popular state-of-the-art baselines on real-world datasets. [ABSTRACT FROM AUTHOR]

Hadoop is an open-source implementation of MapReduce serving for processing large datasets in a massively parallel manner. It was designed aiming at executing large-scale jobs in an enormous number of computing nodes offering computing and storage. However, Hadoop is frequently employed to process short jobs. In practice, short jobs suffer from poor response time and run inefficiently. To fill this gap, this paper analyses the process of job execution and depicts the existing issues why short jobs run inefficiently in Hadoop. According to the characteristic of task execution in multi-wave under cluster overload, we develop a mechanism in light of resource reuse to optimize short jobs execution. This mechanism can reduce the frequency of resource allocation and recovery. Experimental results suggest that the developed mechanism based on resource reuse is able to improve effectiveness of the resource utilization. In addition, the runtime of short jobs can be significantly reduced. [ABSTRACT FROM AUTHOR]

Performance, in terms of quality of service and resource utilization, is one of top attractions in cloud. However, in practice, most multi-tier applications today frequently present large scale fluctuations of response time during periods of high resource utilization. It is hardly to find any reported work that elaborated the reasons causing response time fluctuations, especially when providing quality performance and high effective multi-tier systems in cloud environments. To this end, this paper, through extensive measurements of a multi-tier application benchmark (RUBiS), corroborates the existing of response time fluctuations. In addition, arithmetic mean of response time is not suitable for the measurement of multi-tier services system performance. In light of probing analysis of requests, we show that the large scale response time fluctuations can be caused by concurrent long or mix transactions. The numerical results validate the correctness of our findings. Consequently, an effective scheduling policy called CTP (cross-tier-proportion) is developed to smooth response time fluctuations while maintaining high resource utilization in the system. [ABSTRACT FROM AUTHOR]

Sequential recommendation (SR) provides personalized contents based on the user's historical interactions. Previous SR methods focus on introducing temporal signals of interaction sequence into their sequence encoders without exploring the effect of temporal density information on user preferences. To bridge this gap, we propose the Temporal Density-aware Sequential Recommendation Networks with Contrastive Learning (TDSRec). The specifics of our research mainly consist of two parts. First, we integrate temporal density information into sequential recommendation when capturing user preferences. In detail, through our proposed Temporal KDE Module, we map timestamps into temporal density vectors aiming at improving the recommendation performance. Second, we introduce contrastive learning into TDSRec to alleviate the lack of supervised learning signals as a result of sparse user-item interactions. In detail, instead of directly editing the raw sequence, we leverage two base sequence encoders to derive self-supervision signals. In this way, our approach can naturally eliminate the interference of hand-crafted data augmentation strategies on the raw sequence data. Through extensive experiments on five test datasets, we find that our proposed TDSRec outperforms state-of-the-art baselines. NDCG@10 has been improved by 5.1%, 2.9%, 2.1%, 1.9%, and 1.4% on MovieLens, Beauty, Video Games, CDs&Vinyl, and Movies&TV, respectively. • Integrating temporal kernel density information into sequential recommendation. • Capturing long and short-term user preference via attention and convolution encoders. • Deriving auxiliary self-supervision signals without manual data augmentation. • Extensive experiments with several popular baselines on five real-world datasets. [ABSTRACT FROM AUTHOR]

Abstract: Finite element analyses of tailor-welded blank based on vehicle impact are performed to determine the effects of weld modeling techniques on simulation results. A novel welding model is put forward in this paper, including a simple representation of the weld geometry and material properties, which uses two-fold beams to simulate the behavior of the weld line under impact loading. The normal finite element models of tailor-welded blanks excluding both weld properties and geometry are also established for comparison, in which rigid links and coincident nodes are employed to join the thick and thin parent materials. Simulations of uniaxial tensile tests and impact tests are conducted to study the influences of different weld modeling methods on mechanical behaviors of tailor-welded blanks involved in vehicle impact such as failure position, deformation, and energy absorption. Our results indicate that there are a number of relatively subtle effects associated with the manner in which the weld line is modeled. [Copyright &y& Elsevier]

Abstract: Due to the continuous tightening of various regulations for conserving energy and protecting the environment, the weight reduction of and automobile body becomes an increasingly important issue. In order to reduce the weight of an automobile, we have brought forward a method of replacing the simple structure with the tailor-welded blank (TWB), in which a design process is appropriately defined for a front side rail. Focusing on the thickness determination of TWB, we have deduced the constraint inequalities of strength, bending stiffness, and torsion stiffness to prevent the thickness from decreasing unboundedly due to the lightweight requirement, thus reducing the weight of the front side rail satisfying the design demands of these three factors. The full-width frontal impact simulation of the whole car after redesign in the front side rails has been performed to verify vital performances such as deformation mode, energy absorption, and acceleration. [Copyright &y& Elsevier]

Gossypol, a male antifertility compound isolated from cotton, has been proved to inhibit capacitation and the acrosome reaction in human and mammalian sperm. Here, by using whole-cell recording, we observed the effects of gossypol on Ca2+ and Cl− currents in mouse spermatogenic cells obtained by mechanical dissociation. The results showed that gossypol concentration-dependently and irreversibly inhibited T-type Ca2+ currents in the cells. When the concentration of gossypol was ≥5 μM, the currents were blocked completely. The time to current block was progressively shortened as the gossypol concentration was increased from 5 to 80 μM. Moreover, the drug increased the time constant of inactivation in a concentration-dependent manner, while it did not affect the activation of the current. The inhibitory effect on the T-type Ca2+ current did not correlate with signaling mediated by G proteins and tyrosine phosphorylation. No obvious effect of gossypol on Cl− currents was observed. These data suggest that the gossypol-induced inhibition of T-type Ca2+ currents could be responsible for the antifertility activity of the compound, indicating a possibility to use gossypol as a local contraceptive drug. [Copyright &y& Elsevier]

Pyrolysis characteristics of the corn stalk were investigated at different temperatures and pressures, with focus mainly on the releasing profiles and forming behaviors of gas. The results show that a higher pressure was conducive to the yield of H 2 and CH 4 , and was not conducive to the generation of CO. Combustible gas components increase with the increase temperature. H 2 concentration was affected significantly by temperature, was mainly released at higher temperatures (>500 °C). At 700 °C, the combustible gas components are up to 90% and H 2 concentration is up to 51.78%. The gas components of corn stalk pyrolysis char steam gasification mainly consisted of CO and H 2 , 25.7% and 50.7% respectively. The calorific value of gasification gas is up to 9508.9 kJ/M 3 . A kinetic study of the pyrolysis process of the corn stalk was investigated using a thermogravimetric analyzer. As the heating rate increased, the activity energy reduced. By using the Popescu method to determine that shrinking core model can reflect the pyrolysis process of corn stalk truly at low temperature pyrolysis, and chemical reaction model can reflect the pyrolysis process of corn stalk truly at high temperature pyrolysis. Experimental results showed that values of kinetic parameters can be successfully used to understand the degradation mechanism of solid-state reaction. [ABSTRACT FROM AUTHOR]

Objective: Right‐to‐left ventricle diameter ratio (dRV/dLV) on CT pulmonary angiography (CTPA) is a predictor of outcomes in non‐operated chronic thromboembolic pulmonary hypertension (CTEPH) patients. The purpose of this study is to evaluate the performance of a novel machine learning (ML) algorithm for dRV/dLV measurement in operated CTEPH patients and its association with post‐operative outcomes. Methods: This retrospective study reviewed consecutive CTEPH patients who underwent pulmonary endarterectomy between 2013 and 2017. ML calculated dRV/dLV on pre‐operative CTPA and compared with manual measures. Associations of dRV/dLV with patient characteristics and post‐operative outcomes were evaluated including intensive care (ICU) and hospital length of stay (LOS) using multivariable linear regression analysis. Prolonged LOS was defined as greater than median. Results: ML segmented the ventricles in 99/125 (79%) patients. The most common cause of failure was misidentification of the moderator band as the interventricular septum (7.9%). Mean dRV/dLV by ML was 1.4 ± 0.4 and strongly correlated with manual measures (r = 0.9–0.96 p < 0.0001). dRV/dLV was moderately correlated with measures of pulmonary hypertension on right heart catheterization and RV dilatation on echocardiogram (r = 0.5–0.6, p < 0.0001). dRV/dLV ≥ 1.2 was associated with proximal Jamieson type disease (p = 0.032), longer cardiopulmonary bypass (p = 0.037), aortic cross‐clamp (p = 0.022) and circulatory arrest (p < 0.001) at surgery and dRV/dLV ≥ 1.6 with post‐operative ECMO (p = 0.006). dRV/dLV was independently associated with prolonged ICU LOS (OR = 3.79, 95% CI 1.1–13.06, p = 0.035). Conclusion: dRV/dLV was associated with CTEPH severity and independently associated with prolonged ICU LOS. This CT parameter may therefore assist in perioperative planning. Further refinement of the ML algorithm or CTPA technique is required to avoid errors in ventricular segmentation. Advances in knowledge: Automated right‐to‐left ventricle ratio measurement by machine learning is feasible and is independently associated with outcome after pulmonary endarterectomy. [ABSTRACT FROM AUTHOR]

Objective: Right‐to‐left ventricle diameter ratio (dRV/dLV) on CT pulmonary angiography (CTPA) is a predictor of outcomes in non‐operated chronic thromboembolic pulmonary hypertension (CTEPH) patients. The purpose of this study is to evaluate the performance of a novel machine learning (ML) algorithm for dRV/dLV measurement in operated CTEPH patients and its association with post‐operative outcomes. Methods: This retrospective study reviewed consecutive CTEPH patients who underwent pulmonary endarterectomy between 2013 and 2017. ML calculated dRV/dLV on pre‐operative CTPA and compared with manual measures. Associations of dRV/dLV with patient characteristics and post‐operative outcomes were evaluated including intensive care (ICU) and hospital length of stay (LOS) using multivariable linear regression analysis. Prolonged LOS was defined as greater than median. Results: ML segmented the ventricles in 99/125 (79%) patients. The most common cause of failure was misidentification of the moderator band as the interventricular septum (7.9%). Mean dRV/dLV by ML was 1.4 ± 0.4 and strongly correlated with manual measures (r = 0.9–0.96 p < 0.0001). dRV/dLV was moderately correlated with measures of pulmonary hypertension on right heart catheterization and RV dilatation on echocardiogram (r = 0.5–0.6, p < 0.0001). dRV/dLV ≥ 1.2 was associated with proximal Jamieson type disease (p = 0.032), longer cardiopulmonary bypass (p = 0.037), aortic cross‐clamp (p = 0.022) and circulatory arrest (p < 0.001) at surgery and dRV/dLV ≥ 1.6 with post‐operative ECMO (p = 0.006). dRV/dLV was independently associated with prolonged ICU LOS (OR = 3.79, 95% CI 1.1–13.06, p = 0.035). Conclusion: dRV/dLV was associated with CTEPH severity and independently associated with prolonged ICU LOS. This CT parameter may therefore assist in perioperative planning. Further refinement of the ML algorithm or CTPA technique is required to avoid errors in ventricular segmentation. Advances in knowledge: Automated right‐to‐left ventricle ratio measurement by machine learning is feasible and is independently associated with outcome after pulmonary endarterectomy. [ABSTRACT FROM AUTHOR]

Focuses on a study which elucidated sodium ion channels in human sperm membranes using liposome-reconstituted method. Materials and methods; Results; Discussion.

[Display omitted] • CuCoO x catalyst displayed excellent catalytic performance and water resistance. • Strong interactions between Cu and Co species enhanced humidity VOCs oxidation. • Superb performance was ascribed to high Cu+/Cu2+, Co3+/Co2+ and O ads /O latt ratios. • Type III water adsorbed on CuCoO x facilitated to generate more active oxygen species. • Differences in the mechanism of H 2 O effect on Cu-Co bimetallic oxides were clarified. The impact of water vapor on the catalytic oxidation of VOCs is inevitable in the treatment of industrial waste gases. The presence of water vapor usually inhibits the catalytic oxidation. Therefore, the development of water-resistant, efficient and stable catalysts is of great significance for the control of VOCs. Herein, CuCoO x catalysts were prepared by pyrolysis of in situ synthesized CuCo-ZIF. Meanwhile, a range of Cu-Co bimetallic oxide catalysts were obtained by pyrolysis of Cu/Co-ZIFs synthesized by common strategies for the comparison in the oxidation of aromatic VOCs under humidity conditions. CuCoO x exhibited higher catalytic activity and stability under high moisture conditions compared with CuCoO x -IE (MOF precursors synthesized by ion exchanging) and CuO/Co 3 O 4 (MOF precursors synthesized by mixing). The T 90 of toluene was 209 °C under the reaction conditions of 1000 ppmv toluene, 40,000 mL g−1h−1 WHSV and 5 vol% H 2 O. The improved catalytic performance of CuCoO x was ascribed to more surface lattice defects and enriched active oxygen species, which originated from the enhanced interactions between Cu and Co species. The surface active oxygen species on CuCoO x were either derived from the adsorbed gaseous oxygen, or generated via the reaction of O 2 with type III H 2 O adsorbed on oxygen vacancies. Thus, the presence of H 2 O accelerated the ring-opening of benzoates and attenuated the competitive adsorption of H 2 O and VOCs. This study affords new ideas for the construction of efficient and water-tolerant bimetallic oxide catalysts for the degradation of VOCs under moisture conditions. [ABSTRACT FROM AUTHOR]

Ru-modified monolayer-dispersed WO x /CeO 2 hybrid composites were prepared by co-impregnation (CI) and step impregnation (SI) methods, and the effects of active site distribution on the catalytic oxidation of 1,2-dichloroethane (DCE) were investigated. Ru species tends to be deposited on the monolayer-dispersed WO x (m-WO x) by SI method, which can increase the oxygen vacancies (O V,m-WOx) at m-WO x /CeO 2 interfaces. Abundant O V,m-WOx can promote the formation of more active W–OH and accelerate the dechloridation and oxydehydrogenation of DCE. Oppositely, for CI method, Ru species exists mainly in the form of Ru−O−Ce bonds, increasing the oxygen vacancies (O V,RuCe) into CeO 2 surface lattices and promoting the deep oxidation of intermediate products. The closer contact between the two hetero-interfacial oxygen vacancies (O V,RuCe and O V,m-WOx) on Ru-modified m-WO x /CeO 2 produces a stronger synergistic effect on DCE activation and oxidation, and meanwhile advantageously inhibits the adsorption of chlorine species as well as the formation of polychlorinated by-products. [Display omitted] • Impregnation sequence affects active site distribution on Ru-modified m-WO x /CeO 2 catalysts. • Ru species on m-WO x surfaces increases O V,m-WOx and W–OH by SI method. • CI method generates more Ru−O−Ce and O V,RuCe. • Synergies of hetero-interfacial O V promote DCE activation/oxidation and inhibit Cl species absorption. [ABSTRACT FROM AUTHOR]

• A rotor vibration potato-soil separation method is proposed. • A coupling simulation model of rotor vibration potato-soil separation adopted to DEM-MBD is built. • The optimal parameters of the rotor vibration potato-soil separation device are determined via simulation experiment. • The model and optimal parameters of rotor vibration potato-soil separation is verified via field experiment. A rotor vibration potato-soil separation device (RVPSD) is proposed in view of poor potato-soil separation and higher potato damage rate. Separation efficiency between potatoes and soil and the potato damage rate are selected as evaluation indicators, and a coupling simulation model of potato-soil separation based on Discrete Element Method (DEM) and Multibody Dynamics (MBD) is built up according to structure and working principle of the separation device. The optimal combination of working parameters of the RVPSD is obtained via simulation experiment. The results show that the optimal working parameters of vibration point position, conveying speed of potato-soil separation elevating chain, rotor amplitude and rotor vibration frequency are 646.5 mm, 1.08 m/s, 26.7 mm and 5.9 Hz respectively. The field validation experiment is carried out based on the optimal combination parameters. The results show that the potato-soil separation efficiency and potato damage rate of the RVPSD are 97.8 % and 1.16 % respectively, the field experiment results are basically consistent with the simulation results, which proves the correctness of the simulation model. It can provide theoretical reference for rotor vibration potato-soil separation process simulation and device parameter optimization. [ABSTRACT FROM AUTHOR]

• A cucumber canopy vine top recognition algorithm based on YOLOv5 is proposed. • The cucumber canopy vine top recognition performance of improved algorithm is analyzed and compared. • The recognition performance of the improved algorithm is verified via instance recognition. • The improved algorithm has good recognition performance for the cucumber canopy vine tops. In order to solve the problem of low recognition accuracy of cucumber canopy vine tops image caused by the high density and intermingling of the cucumber canopy vine, an image recognition model of cucumber canopy tops based on the deep learning model of improved YOLOv5 is proposed in this paper. A combination of online and offline augmentation is used to amplify the original data to ensure a balanced number of images in different types of datasets during training and testing. Considering the small leaves characteristics of the cucumber canopy vine tops, the CA mechanism module (YOLOv5-CA) is introduced into the backbone network of the YOLOv5 model to improve the recognition accuracy of the small leaf target. On this basis, the regression Loss GIOU of the target position was changed to EIOU and the cross-entropy function of confidence loss was changed to Focal Loss function (YOLOv5-CA-LS) to reduce the influence of the imbalance of difficult and easy samples on the detection results. The models before and after optimization were compared. The results show that compared with the YOLOv5 model, the accuracy, recall rate, mAP@0.5 and mAP@0.5:0.95 of the YOLOv5-CA-LS model is improved by 4.4%, 5.0%, 1.3% and 1.8% respectively. And average recognition accuracy value of the instance of the cucumber canopy tops adopted the YOLOv5-CA-LS model is 97.1% under different angle and cloudy conditions, which can meet the recognition requirement of high density and intermingling of the cucumber canopy vine tops. [ABSTRACT FROM AUTHOR]

Hashing based approximate nearest neighbor (ANN) search techniques have attracted considerable attention in media search community because of its good potential for low storage cost and fast query speed. Many hashing based ANN search methods have been proposed; but, most of them just consider to keep the similarity relationship of data points during mapping instead of topology of data. It is well known that Self-Organizing Map can keep topology structure while conducting mapping task. Motivated by this, in this paper, we propose a Self-Organizing Map based hashing framework–SOMH, which cannot only keep similarity relationship, but also preserve topology of data. Specifically, in SOMH, Self-Organizing Map is introduced to map data points into hamming space. In addition, in order to make it work well on short and long binary codes, we propose a relaxed version of SOMH and a product space SOMH, respectively. For the optimization problem of the relaxed SOMH, we also present an iterative solution. Moreover, we further propose an extended version of SOMH, which can work well on multimodal data search task, i.e., cross-modal search. To test the performance of these proposed algorithms, we conduct experiments on three data sets–SIFT1M, GIST1M and Wiki (a multimodal dataset). Experimental results show that SOMH can outperform or is comparable to several state-of-the-arts. [ABSTRACT FROM AUTHOR]

• The BFRC triaxial properties was effected by the fiber orientation and content. • Excessive fiber can lead to increased porosity, the primary defect of concrete. • The fibers were pulled out rather than broken when BFRC failure. • CT technology will be a powerful tool for studying the internal structure of BFRC. • The low density fibers in CT model can be identified by the search cone algorithm. The addition of basalt fiber can effectively improve the macroscopic performance of concrete. At present, there are few studies on the micro-action mechanism of basalt fiber. In this paper, the triaxial mechanical properties of basalt fibers reinforced concrete (BFRC) are investigated, and the internal microstructure of BFRC is studied with the aid of X-ray Computerized tomography (CT technology), digital image processing technology (DIP) and SEM to study the reinforcement mechanism of basalt fibers on concrete from a microscopic point of view. The study showed that the peak stresses under the same peri-pressure showed an increasing and then decreasing trend with the fiber dosage, and the optimum fiber dosage was 3 kg/m3; the excessive addition of fibers led to the increase of internal pores, the contact probability increased, and the peak stresses decreased. The influence to the triaxial mechanical properties of BFRCs were ranked by the fiber dosage, lesser extent and fiber orientation. In this paper, the study of basalt fiber toughness enhancement from both concrete matrix defects and fiber distribution can further advance the application of ductile concrete composites research. [ABSTRACT FROM AUTHOR]