Search

To solve the tracking accuracy degradation problem in scenarios with large displacements or nonrigid motion during target tracking, this paper proposes an optical flow estimation method based on multiscale anisotropic convolution. The network structure is improved in a step-by-step manner by extracting the data flow from the network according to the observed features. For the low-level neural network, a layered multiscale structure is used to build a cascade network by using hybrid dilated convolution to obtain feature information at different scales while ensuring the tracking accuracy. For the upper-layer neural network, hybrid inflated deformable convolution is used by learning the contextual long-range correlations and multidirectional adaptive offsets of features. Experiments are conducted on the Flying Chairs, KITTI, and MPI datasets. The results show that compared with various popular algorithm methods, the model in this paper reduces endpoint errors while retaining edge information in regions with large displacements or nonrigid motion. Code is available at https://github.com/yifanna/MACFlow-pytorch. [ABSTRACT FROM AUTHOR]

Fine-grained management of rice fields can enhance the yield and quality of rice crops. Challenges in achieving fine classification include interference from similar vegetation, the irregularity of natural field shapes, and complex scale variations. This paper introduces Rice Attention Cascade Network (RACNet), for the fine classification of rice fields in high-resolution satellite remote sensing imagery. The network employs the Hybrid Task Cascade network as the base framework and uses spectral and indices mixed multimodal data as input to reinforce the feature differentiation of similar vegetation. Initially, a Channel Attention Deformable-ResNet (CAD-ResNet) was designed to enhance the feature representation of rice on different channels. Deformable convolution improves the ability of CAD-ResNet to capture irregular field shapes. Then, to address the issue of complex scale changes, the multi-scale features extracted by the CAD-ResNet are progressively fused using an Asymptotic Feature Pyramid, reducing the loss of scale information between non-adjacent layers. Experiments on the Meishan rice dataset show that the proposed method is capable of accurate instance segmentation for fragmented or irregularly shaped rice fields. The evaluation metric AP50 of RACNet reaches 50.8%. [ABSTRACT FROM AUTHOR]

Defect detection in complex background is a critical issue. To address this issue, this paper proposes the mixture attention mechanism cascade network, in which the new channel attention network is linked with the spatial attention network to create an effective mixed attention network that takes advantage of their respective advantages, adaptively suppresses background noise, and highlights defect features. To ensure the efficiency and effectiveness of effective mixed attention network, the new channel attention network splices the output features of the global average pooling layer and the global maximum pooling layer and then sends the spliced features into a shared network, which is a one‐dimensional convolutional network, and uses cross‐channel interaction for fusion. Furthermore, in order to provide more discriminative feature representation, the authors extract the intermediate features of the region proposal network and input them into effective mixed attention network. Finally, the cascade head is used to refine the predicted bounding box to achieve high‐quality defect location. To demonstrate the superiority and usefulness of this method, it is compared to the latest method using widely used PCB and NEU data sets. A large number of trials demonstrate that this strategy outperforms other methods for detecting defects in complicated backgrounds. [ABSTRACT FROM AUTHOR]

The rise of social networks has greatly contributed to creating information cascades. Overtime, new nodes are added to the cascade network, which means the cascade network is dynamically variable. At the same time, there are often only a few nodes in the cascade network before new nodes join. Therefore, it becomes a key task to predict the diffusion after the dynamic cascade based on the small number of nodes observed in the previous period. However, existing methods are limited for dynamic short cascades and cannot combine temporal information with structural information well, so a new model, MetaCaFormer, based on meta-learning and the Transformer structure, is proposed in this paper for dynamic short cascade prediction. Considering the limited processing capability of traditional graph neural networks for temporal information, we propose a CaFormer model based on the Transformer structure, which inherits the powerful processing capability of Transformer for temporal information, while considering the neighboring nodes, edges and spatial importance of nodes, effectively combining temporal and structural information. At the same time, to improve the prediction ability for short cascades, we also fuse meta-learning so that it can be quickly adapted to short cascade data. In this paper, MetaCaFormer is applied to two publicly available datasets in different scenarios for experiments to demonstrate its effectiveness and generalization ability. The experimental results show that MetaCaFormer outperforms the currently available baseline methods. [ABSTRACT FROM AUTHOR]

Cracks are one of the main types of tunnel-lining defects. At present, there are no particularly good methods for identifying tunnel-lining cracks. The methods that are used are associated with problems such as poor robustness, low detection efficiency, and inconsistency in defect identification. Vision-based crack identification algorithms that use images of tunnel-lining cracks are also affected by problems such as weak features, irregular shapes, random development, and small proportions. In light of this, we propose in this paper a neural network that integrates U-net and FPN (feature pyramid network) pyramid structure characteristics. Because of its ability to output deep feature information, an SK (selective kernel) attention mechanism was integrated to increase the weight of effective feature regions and improve the detection accuracy of feature maps. An RPN (region proposal network) regional detection network was used for multi-scale regional screening, and non-maximum suppression was used to eliminate repeated anchor frames between different regions. The continuity of cracks was determined using adaptive regional expansion. The region segmentation network was also combined with multi-scale feature maps. Finally, the segmentation results for regions of fine features could be mapped to the whole image to identify cracks, thereby solving the problem that arises when textural features of cracks are weak and cannot be accurately identified. For testing purposes, photos of tunnel-lining crack defects were used. The algorithm described in this paper was able to combine deep and shallow features to identify more abundant crack features, and the recognition accuracy of the crack classification network reached 98.51%. With respect to crack segmentation, the algorithm's segmentation accuracy reached 94.55%, and its single processing time reached 60.76 ms, indicating more accurate and efficient segmentation performance, compared with FCNs (fully convolutional networks) and classical U-net networks. [ABSTRACT FROM AUTHOR]

Construction collapse accidents are one of the most serious accidents. A variety of causal factors interact with each other, leading to the occurrence of accidents. Some causal factors may cause other causal factors to happen one after another, eventually leading to the occurrence of accidents. The propagation process of causal factors that can lead to construction collapse accident must be clarified. Therefore, this study aims to analyze the propagation path of construction collapse causal factors based on network theory and cascading failure model. First, this study collected and screened the cases of construction collapse accidents, and conducted a statistical analysis of their basic information. Second, a construction collapse accident causation (CCAC) framework was established to identify causal factors in accident cases. Then, a CCAC network model was established. Based on multiple measurement criteria in complex network theory, the causal factors were ranked. After that, through the cascading failure analysis of the causal factors, the critical causal factor-accident propagation path was obtained. It is found that construction scheme and organization is at the beginning and is the source of other causal factors. Finally, five cases were analyzed to verify the rationality of the results. This paper provides a solution to study the propagation path of causal factors and find out the characteristics of the CCAC mechanism through a large number of accident cases, which has theoretical and practical significance for the prevention of construction collapse accidents. [ABSTRACT FROM AUTHOR]

A safe and stable operation power system is very important for the maintenance of national industrial security and social economy. However, with the increasing complexity of the power grid topology and its operation, new challenges in estimating and evaluating the grid structure performance have received significant attention. Complex network theory transfers the power grid to a network with nodes and links, which helps evaluate the system conveniently with a global view. In this paper, we employ the complex network method to address the cascade failure process and grid structure performance assessment simultaneously. Firstly, a grid cascade failure model based on network topology and power system characteristics is constructed. Then, a set of performance evaluation indicators, including invulnerability, reliability, and vulnerability, is proposed based on the actual functional properties of the grid by renewing the power-weighted degree, medium, and clustering coefficients according to the network cascade failure. Finally, a comprehensive network performance evaluation index, which combines the invulnerability, reliability, and vulnerability indicators with an entropy-based objective weighting method, is put forward in this study. In order to confirm the approach's efficacy, an IEEE-30 bus system is employed for a case study. Numerical results show that the weighted integrated index with a functional network could better evaluate the power grid performance than the unweighted index with a topology network, which demonstrates and validates the effectiveness of the method proposed in this paper. [ABSTRACT FROM AUTHOR]

One of the most challenging issues in contemporary complex network research is to understand the structure and vulnerability of multilayer networks, even though cascading failures in single networks have been widely studied in recent years. The goal of this work is to compare the similarities and differences between four single layers and understand the implications of interdependencies among cities on the overall vulnerability of a multilayer global logistics network. In this paper, a global logistics network model set as a multilayer network considering cascading failures is proposed in different disruption scenarios. Two types of attack strategies—a highest load attack and a lowest load attack—are used to evaluate the vulnerability of the global logistics network and to further analyze the changes in the topology properties. For a multilayer network, the vulnerability of single layers is compared as well. The results suggest that compared with the results of a single global logistics network, a multilayer network has a higher vulnerability. In addition, the heterogeneity of networks plays an important role in the vulnerability of a multilayer network against targeted attacks. Protecting the most important nodes is critical to safeguard the potential "vulnerability" in the development of the global logistics network. The three-step response strategy of "Prewarning–Response–Postrepair" is the main pathway to improving the adjustment ability and adaptability of logistics hub cities in response to external shocks. These findings supplement and extend the previous attack results on nodes and can thus help us better explain the vulnerability of different networks and provide insight into more tolerant, real, complex system designs. [ABSTRACT FROM AUTHOR]

The design level of channel physical characteristics has a crucial influence on the transmission quality of high-speed serial links. However, channel design requires a complex simulation and verification process. In this paper, a cascade neural network model constructed of a Deep Neural Network (DNN) and a Transformer is proposed. This model takes physical features as inputs and imports a Single-Bit Response (SBR) as a connection, which is enhanced through predicting frequency characteristics and equalizer parameters. At the same time, signal integrity (SI) analysis and link optimization are achieved by predicting eye diagrams and channel operating margins (COMs). Additionally, Bayesian optimization based on the Gaussian process (GP) is employed for hyperparameter optimization (HPO). The results show that the DNN–Transformer cascaded model achieves high-precision predictions of multiple metrics in performance prediction and optimization, and the maximum relative error of the test-set results is less than 2% under the equalizer architecture of a 3-taps TX FFE, an RX CTLE with dual DC gain, and a 12-taps RX DFE, which is more powerful than other deep learning models in terms of prediction ability. [ABSTRACT FROM AUTHOR]

The necessary and sufficient conditions for decomposing a fourth-order (FO) linear time-variant system (LTVS) in the form of cascade connection of a first-order system and a third-order system that are commutative are expressed explicitly in this paper. Some additional conditions are also derived for a valid decomposition when the initial conditions are different from zero. Further, explicit formulas are derived for the determination of the decomposed subsystems. The paper also investigated the effects of changing the connection order on the sensitivity and disturbance properties. Theoretical results about the decomposition are supported and illustrated by considering various examples. [ABSTRACT FROM AUTHOR]

This paper focuses on the problem of the modeling of FET power transistors made of gallium nitride offered by GaN Systems, Transphorm, and Nexperia. The considered devices have been available on the market since 2014. GaN-FETs are built as a cascade connection of a normally on gallium nitride HEMT and a normally off MOSFET made of silicon. On the manufacturer's sites, one can find models of these devices for like-SPICE tools in the text form. The main goal of this paper is to evaluate the model's accuracy by comparing calculation results obtained by the use of the considered models with the authors' measurement results and datasheet. It has been demonstrated that the GaN Systems model built on controlled sources described by a set of arbitrarily selected mathematical functions more accurately reproduces the basic characteristics of a transistor. On the other hand, the models from Transphorm and Nexperia, which are constructed based on built-in semiconductor device models, more precisely calculate the values of selected functional transistor parameters. [ABSTRACT FROM AUTHOR]

Inverse synthetic aperture LiDAR (ISAL) can create high-resolution images within a few milliseconds, which are employed for long-range airspace target identification. However, its optical signal characteristics incur the non-negligible higher-order kinematic parameters of the target and phase errors due to atmospheric turbulence. These higher-order parameters and phase errors make it challenging for imaging the ISAL signals. In this paper, we propose an approach integrating the RD algorithm with an image translation network. Unlike the conventional methods, our approach does not require high accuracy in estimating each target motion and atmospheric parameter. The phase error of the RD image is fitted by an image translation network, which greatly simplifies the computational difficulty of the ISAL imaging model. The experimental results demonstrate that our model has good generalization performance. Specifically, our method consistently performs well in capturing the target information under different types of noise and sparsity aperture (SA) rates compared to other conventional methods. In addition, our approach can be applied to the measured data after training the network by using simulated data. [ABSTRACT FROM AUTHOR]

Many real-world networks with Scale-Free structure are significantly vulnerable against both intentional attacks and catastrophic cascading failures. On the other hand, it has been shown that networks with narrower degree distributions have strong robustness of connectivity by enhancing loops. This paper numerically reveals that such networks are also tolerant against cascading failures. Our findings will be useful in designing stronger tolerant network infrastructures. [ABSTRACT FROM AUTHOR]

This paper identifies and analyzes the microstructure of a carburized layer by using a deep convolutional neural network, selecting different carburizing processes to conduct surface treatment on 23CrNi3Mo steel, collecting many metallographic pictures of the carburized layer based on laser confocal microscopy, and building a microstructure dataset (MCLD) database for training and testing. Five algorithms—a full convolutional network (FCN), U-Net, DeepLabv3+, pyramid scene parsing network (PSPNet), and image cascade network (ICNet)—are used to segment the self-built microstructural dataset (MCLD). By comparing the five deep learning algorithms, a neural network model suitable for the MCLD database is identified and optimized. The research results achieve recognition, segmentation, and statistic verification of metallographic microstructure images through a deep convolutional neural network. This approach can replace the high cost and complicated process of experimental testing of retained austenite and martensite. This new method is provided to identify and calculate the content of residual austenite and martensite in the carburized layer of low-carbon steel, which lays a theoretical foundation for optimizing the carburizing process. [ABSTRACT FROM AUTHOR]

In recent years, the development of deep learning has brought great convenience to the work of target detection, semantic segmentation, and object recognition. In the field of infrared weak small target detection (e.g., surveillance and reconnaissance), it is not only necessary to accurately detect targets but also to perform precise segmentation and sub-pixel-level centroid localization for infrared small targets with low signal-to-noise ratio and weak texture information. To address these issues, we propose UCDnet (Double U-shaped Segmentation Network Cascade Centroid Map Prediction for Infrared Weak Small Target Detection) in this paper, which completes "end-to-end" training and prediction by cascading the centroid localization subnet with the semantic segmentation subnet. We propose the novel double U-shaped feature extraction network for point target fine segmentation. We propose the concept and method of centroid map prediction for point target localization and design the corresponding Com loss function, together with a new centroid localization evaluation metrics. The experiments show that ours achieves target detection, semantic segmentation, and sub-pixel-level centroid localization. When the target signal-to-noise ratio is greater than 0.4, the IoU of our semantic segmentation results can reach 0.9186, and the average centroid localization precision can reach 0.3371 pixels. On our simulated dataset of infrared weak small targets, the algorithm we proposed performs better than existing state-of-the-art networks in terms of semantic segmentation and centroid localization. [ABSTRACT FROM AUTHOR]

Fruit tree diseases are one of the major agricultural disasters in China. With the popularity of smartphones, there is a trend to use mobile devices to identify agricultural pests and diseases. In order to identify leaf diseases of apples more easily and efficiently, this paper proposes a cascade backbone network-based (CBNet) disease identification method to detect leaf diseases of apple trees in the field. The method first replaces traditional convolutional blocks with MobileViT-based convolutional blocks particularly for feature extraction. Compared with the traditional convolutional block, the MobileViT-based convolutional block is able to mine feature information in the image better. In order to refine the mined feature information, a feature refinement module is proposed in this paper. At the same time, this paper proposes a cascaded backbone network for effective fusion of features using a pyramidal cascaded multiplication operation. The results conducted on field datasets collected using mobile devices showed that the network proposed in this paper can achieve 96.76% accuracy and 96.71% F1-score. To the best of our knowledge, this paper is the first to introduce Transformer into apple leaf disease identification, and the results are promising. [ABSTRACT FROM AUTHOR]

Eddies of various sizes are visible to the naked eye in turbulent flow. Each eddy scale corresponds to a fraction of the total energy released by the turbulence cascade. Understanding the dynamic mechanism of the energy cascade is crucial to the study of turbulent mixing. In this paper, an energy cascade multi-layer network (ECMN) based on the complex network algorithm is proposed to investigate the spatio-temporal evolution of the energy cascade, covering both the inertial and dispersive ranges. The dynamic process of energy cascade is transformed into a topological structure based on the node definition and edge determination. The topological structure allows for the exploration of eddies interaction and chaotic energy transfer across scales. The model results show the intermittent and non-uniform nature of the energy cascade. Meanwhile, the scale gap found in the model verifies the fractal property of the energy evolution. We also found that scales of the generated eddies in energy cascade process are stochastic, and a synchronous energy cascade pattern is demonstrated according to the constructed framework. Furthermore, it provides a topological way to evaluate the contribution of large and small scale eddies. In addition, a network structure coefficient κ is proposed to evaluate the energy transfer strength. It agrees very well with the fluctuation of dissipation rates. All of this shows that the network model can effectively reveal the inhomogeneous properties of the energy cascade and quantify the turbulent mixing intensity based on the intermittent scale interaction. This also provides new insights into the study of fractal scales of nonlinear complex systems and the bridging of chaotic dynamics with topological frameworks. [ABSTRACT FROM AUTHOR]

According to experts in neurology, brain tumours pose a serious risk to human health. The clinical identification and treatment of brain tumours rely heavily on accurate segmentation. The varied sizes, forms, and locations of brain tumours make accurate automated segmentation a formidable obstacle in the field of neuroscience. U-Net, with its computational intelligence and concise design, has lately been the go-to model for fixing medical picture segmentation issues. Problems with restricted local receptive fields, lost spatial information, and inadequate contextual information are still plaguing artificial intelligence. A convolutional neural network (CNN) and a Mel-spectrogram are the basis of this cough recognition technique. First, we combine the voice in a variety of intricate settings and improve the audio data. After that, we preprocess the data to make sure its length is consistent and create a Mel-spectrogram out of it. A novel model for brain tumor segmentation (BTS), Intelligence Cascade U-Net (ICUNet), is proposed to address these issues. It is built on dynamic convolution and uses a non-local attention mechanism. In order to reconstruct more detailed spatial information on brain tumours, the principal design is a two-stage cascade of 3DU-Net. The paper's objective is to identify the best learnable parameters that will maximize the likelihood of the data. After the network's ability to gather long-distance dependencies for AI, Expectation--Maximization is applied to the cascade network's lateral connections, enabling it to leverage contextual data more effectively. Lastly, to enhance the network's ability to capture local characteristics, dynamic convolutions with local adaptive capabilities are used in place of the cascade network's standard convolutions. We compared our results to those of other typical methods and ran extensive testing utilising the publicly available BraTS 2019/2020 datasets. The suggested method performs well on tasks involving BTS, according to the experimental data. The Dice scores for tumor core (TC), complete tumor, and enhanced tumor segmentation BraTS 2019/2020 validation sets are 0.897/0.903, 0.826/0.828, and 0.781/0.786, respectively, indicating high performance in BTS. [ABSTRACT FROM AUTHOR]

With the development of the Internet, vast amounts of text information are being generated constantly. Methods for extracting the valuable parts from this information have become an important research field. Relation extraction aims to identify entities and the relations between them from text, helping computers better understand textual information. Currently, the field of relation extraction faces various challenges, particularly in addressing the relation overlapping problem. The main difficulties are as follows: (1) Traditional methods of relation extraction have limitations and lack the ability to handle the relation overlapping problem, requiring a redesign. (2) Relation extraction models are easily disturbed by noise from words with weak relevance to the relation extraction task, leading to difficulties in correctly identifying entities and their relations. In this paper, we propose the Relation extraction method based on the Entity Attention network and Cascade binary Tagging framework (REACT). We decompose the relation extraction task into two subtasks: head entity identification and tail entity and relation identification. REACT first identifies the head entity and then identifies all possible tail entities that can be paired with the head entity, as well as all possible relations. With this architecture, the model can handle the relation overlapping problem. In order to reduce the interference of words in the text that are not related to the head entity or relation extraction task and improve the accuracy of identifying the tail entities and relations, we designed an entity attention network. To demonstrate the effectiveness of REACT, we construct a high-quality Chinese dataset and conduct a large number of experiments on this dataset. The experimental results fully confirm the effectiveness of REACT, showing its significant advantages in handling the relation overlapping problem compared to current other methods. [ABSTRACT FROM AUTHOR]

Recently, lung cancer prediction based on imaging genomics has attracted great attention. However, such studies often have many challenges, such as small sample size, high-dimensional information redundancy, and the inefficiency of multimodal fusion. Therefore, in this paper, a deep convolution cascade attention fusion network (DCCAFN) based on imaging genomics is proposed for the prediction of lung cancer patients' survival. The network consists of three modules: an image feature extraction module (IFEM), a gene feature extraction module (GFEM), and an attention fusion network (AFN). In the IFEM, a pretrained residual network based on transfer learning is used to extract deep image features to fully capture the computed tomography (CT) image information conducive to prognosis prediction. In the GFEM, the F-test is first used for gene screening to eliminate redundant information, and then, a cascade network with the convolution cascade module (CCM) that contains a convolution operation, a pooling operation, and an ensemble forest classifier is designed to better extract the gene features. In the AFN, a bimodal attention fusion mechanism is proposed to fuse deep image features and gene features to improve the performance of predicting lung cancer survival. The experimental results show that the DCCAFN model achieves good performance, and its accuracy and AUC are 0.831 and 0.816, respectively. It indicates that the model is an effective multimodal data fusion method for predicting the survival prognosis of lung cancer, which can greatly help physicians stratify patients' risks, and achieve personalized treatment for improving the quality of patients' lives. [ABSTRACT FROM AUTHOR]

This paper presents a unified model for combining beamforming and blind source separation (BSS). The validity of the model's assumptions is confirmed by recovering target speech information in noise accurately using Oracle information. Using real static human-robot interaction (HRI) data, the proposed combination of BSS with the minimum-variance distortionless response beamformer provides a greater signal-to-noise ratio (SNR) than previous parallel and cascade systems that combine BSS and beamforming. In the difficult-to-model HRI dynamic environment, the system provides a SNR gain that was 2.8 dB greater than the results obtained with the cascade combination, where the parallel combination is infeasible. [ABSTRACT FROM AUTHOR]

Knowing how to effectively predict the scale of future information cascades based on the historical trajectory of information dissemination has become an important topic. It is significant for public opinion guidance; advertising; and hotspot recommendation. Deep learning technology has become a research hotspot in popularity prediction, but for complex social platform data, existing methods are challenging to utilize cascade information effectively. This paper proposes a novel end-to-end deep learning network CAC-G with cascade attention convolution (CAC). This model can stress the global information when learning node information and reducing errors caused by information loss. Moreover, a novel Dynamic routing-AT aggregation method is investigated and applied to aggregate node information to generate a representation of cascade snapshots. Then, the gated recurrent unit (GRU) is employed to learn temporal information. This study's validity and generalization ability are verified in the experiments by applying CAC-G on two public datasets where CAC-G is better than the existing baseline methods. [ABSTRACT FROM AUTHOR]

A reasonable definition of nodes load and capacity is essential for improving the robustness of scale-free networks against cascading failure, which has gained significant attention over recent years. This paper presents two methods for defining the load-capacity model: a degree-based method and a betweenness-based method. In these methods, the initial load and capacity of nodes were determined by considering the degrees and betweenness centrality of nodes and their neighbors. These values could be adjusted using both global and local parameters. This paper achieved load redistribution during cascading failures through targeted attacks on network nodes. In addition, this study applied load redistribution to cascading failure processes in networks by targeting network nodes. In order to evaluate the effectiveness of the proposed approach, this paper examines the impact of adjusting two parameters on the minimum critical tolerance coefficient and network robustness. Computer-generated scale-free networks and a real network were used for evaluation purposes. The findings indicated that higher global parameters resulted in a lower average robustness index. Moreover, our degree-based method demonstrated a smaller minimum critical tolerance coefficient and average robustness index compared to existing load definition methods. Therefore, the proposed methods enhanced the robustness and integrity of scale-free networks against attacks. [ABSTRACT FROM AUTHOR]

In this paper, a fast prediction model of the electromagnetic response of the LTE-R (Long Term Evolution for Railway) communication system based on a cascade neural network is developed to quickly analyze the impact of pantograph arcing on the LTE-R system during vehicle operation. In order to obtain the coupling disturbance level of the LTE-R antenna port, this model uses the cascade neural network based on the PSO-BP (Particle Swarm Optimization of Back Propagation Neural Network) algorithm to quickly solve the coupling coefficient of the pantograph arcing measurement probe and the antenna port. A two-stage BP neural network model is used to train both the simulation data and measurement data, and the results are validated by field tests. Finally, the antenna port coupling interference is added to the LTE-R system to analyze the impact of pantograph arcing on the quality of communication transmission. Analysis and tests demonstrate that pantograph arcing can lead to an increase in the BLER of LTE-R systems at 400 MHz, affecting system performance and transmission efficiency. [ABSTRACT FROM AUTHOR]

The article presents the use of the Texas Instruments LM5170EVM-BIDIR bidirectional DC/DC converter to control power distribution in a hybrid energy storage system based on a battery–ultracapacitor system. The paper describes typical topologies of connecting a battery with an ultracapacitor. The results of tests for calibration and identification of converter parameters are presented. The main innovation of the solution presented in this paper is the appropriate selection of the nominal voltage of the ultracapacitor so that the converter can be operated only in the constant current mode, in a cascade connection, excluding the low-efficiency constant voltage mode. This article demonstrated that such control allows for high efficiency and reduction of losses in the DC/DC converter, which is necessary in the case of mobile solutions. The amount of losses was determined depending on the control voltage in the operation modes of the converter: in the Step Up mode by increasing the voltage from 12 V to 24 V, from 12 V to 36 V, and from 12 V to 48 V and in the Step Down mode by decreasing the voltage from 48 V to 12 V, from 36 V to 12 V, and from 24 V to 12 V. For a calibrated converter in a semi-active topology, bench tests were carried out in a cycle with pulsating load. The tests were carried out using LiFePO4 cells with a voltage of 12 V and Maxwell ultracapacitors with a package voltage of 48 V. Power distribution in the range of 10% to 90% was achieved using the myRIO platform, which controlled the operation of the DC/DC converter based on an external current profile. [ABSTRACT FROM AUTHOR]

Crack detection plays a critical role in ensuring road safety and maintenance. Traditional, manual, and semi-automatic detection methods have proven inefficient. Nowadays, the emergence of deep learning techniques has opened up new possibilities for automatic crack detection. However, there are few methods with both localization and segmentation abilities, and most perform poorly. The consistent nature of pavement over a small mileage range gives us the opportunity to make improvements. A novel data-augmentation strategy called CrackMover, specifically tailored for crack detection methods, is proposed. Experiments demonstrate the effectiveness of CrackMover for various methods. Moreover, this paper presents a new instance segmentation method for crack detection. It adopts a redesigned backbone network and incorporates a cascade structure for the region-based convolutional network (R-CNN) part. The experimental evaluation showcases significant performance improvements achieved by these approaches in crack detection. The proposed method achieves an average precision of 33.3%, surpassing Mask R-CNN with a Residual Network 50 backbone by 8.6%, proving its effectiveness in detecting crack distress. [ABSTRACT FROM AUTHOR]

Recently, the field of vehicle-mounted visual intelligence technology has witnessed a surge of interest in pedestrian detection. Existing algorithms for dense pedestrian detection at intersections face challenges such as high computational weight, complex models that are difficult to deploy, and suboptimal detection accuracy for small targets and highly occluded pedestrians. To address these issues, this paper proposes an improved lightweight multi-scale pedestrian detection algorithm, YOLOv8-CB. The algorithm introduces a lightweight cascade fusion network, CFNet (cascade fusion network), and a CBAM attention module to improve the characterization of multi-scale feature semantics and location information, and it superimposes a bidirectional weighted feature fusion path BIFPN structure to fuse more effective features and improve pedestrian detection performance. It is experimentally verified that compared with the YOLOv8n algorithm, the accuracy of the improved model is increased by 2.4%, the number of model parameters is reduced by 6.45%, and the computational load is reduced by 6.74%. The inference time for a single image is 10.8 ms. The cascade fusion algorithm YOLOv8-CB has higher detection accuracy and is a lighter model for multi-scale pedestrian detection in complex scenes such as streets or intersections. This proposed algorithm presents a valuable approach for device-side pedestrian detection with limited computational resources. [ABSTRACT FROM AUTHOR]

One of the main disadvantages of the traditional mean square error (MSE)-based constructive networks is their poor performance in the presence of non-Gaussian noises. In this paper, we propose a new incremental constructive network based on the correntropy objective function (correntropy-based constructive neural network (C2N2)), which is robust to non-Gaussian noises. In the proposed learning method, input and output side optimizations are separated. It is proved theoretically that the new hidden node, which is obtained from the input side optimization problem, is not orthogonal to the residual error function. Regarding this fact, it is proved that the correntropy of the residual error converges to its optimum value. During the training process, the weighted linear least square problem is iteratively applied to update the parameters of the newly added node. Experiments on both synthetic and benchmark datasets demonstrate the robustness of the proposed method in comparison with the MSE-based constructive network, the radial basis function (RBF) network. Moreover, the proposed method outperforms other robust learning methods including the cascade correntropy network (CCOEN), Multi-Layer Perceptron based on the Minimum Error Entropy objective function (MLPMEE), Multi-Layer Perceptron based on the correntropy objective function (MLPMCC) and the Robust Least Square Support Vector Machine (RLS-SVM). [ABSTRACT FROM AUTHOR]

The material network in the chemical process has the characteristics of a complex network. Once the fault occurs, it will spread and lead to cascading failures, causing the chemical production process to not proceed normally. Previous studies of cascading failures have been limited by the size and characteristics of the network, which prevented the definition of the fault propagation capacity of the edge based on the actual situation. Studies on the effects of network topology and certain risk factors on the amount of fault propagation still need to be completed. Because of the above shortcomings, this study constructs a chemical material network cascading failure invulnerability analysis model that considers the failure propagation ability. Through the study of the chemical material network topology structure and the analysis of various flammable and explosive chemicals, the paper defines the fault propagation coefficient, which affects the load propagation quantity on the connecting edge. Then, based on the load‐capacity nonlinear cascading failure model, the chemical material network's invulnerability was analyzed by adjusting the model parameters. The effectiveness of the model is verified through case analysis. Compared to the existing model, the model in this paper effectively improves the invulnerability of the material network and can reduce the phenomenon of large‐scale failure of the material network due to the failure propagation capability is not considered. [ABSTRACT FROM AUTHOR]

Image-based gender classification is very useful in many applications, such as intelligent surveillance, micromarketing, etc. One common approach is to adopt a machine learning algorithm to recognize the gender class of the captured subject based on spatio-temporal gait features extracted from the image. The image input can be generated from the video of the walking cycle, e.g., gait energy image (GEI). Recognition accuracy depends on the similarity of intra-class GEIs, as well as the dissimilarity of inter-class GEIs. However, we observe that, at some viewing angles, the GEIs of both gender classes are very similar. Moreover, the GEI does not exhibit a clear appearance of posture. We postulate that distinctive postures of the walking cycle can provide additional and valuable information for gender classification. This paper proposes a gender classification framework that exploits multiple inputs of the GEI and the characteristic poses of the walking cycle. The proposed framework is a cascade network that is capable of gradually learning the gait features from images acquired in multiple views. The cascade network contains a feature extractor and gender classifier. The multi-stream feature extractor network is trained to extract features from the multiple input images. Features are then fed to the classifier network, which is trained with ensemble learning. We evaluate and compare the performance of our proposed framework with state-of-the-art gait-based gender classification methods on benchmark datasets. The proposed framework outperforms other methods that only utilize a single input of the GEI or pose. [ABSTRACT FROM AUTHOR]

Table detection is a preliminary step in extracting reliable information from tables in scanned document images. We present CasTabDetectoRS, a novel end-to-end trainable table detection framework that operates on Cascade Mask R-CNN, including Recursive Feature Pyramid network and Switchable Atrous Convolution in the existing backbone architecture. By utilizing a comparativelyightweight backbone of ResNet-50, this paper demonstrates that superior results are attainable without relying on pre- and post-processing methods, heavier backbone networks (ResNet-101, ResNeXt-152), and memory-intensive deformable convolutions. We evaluate the proposed approach on five different publicly available table detection datasets. Our CasTabDetectoRS outperforms the previous state-of-the-art results on four datasets (ICDAR-19, TableBank, UNLV, and Marmot) and accomplishes comparable results on ICDAR-17 POD. Upon comparing with previous state-of-the-art results, we obtain a significant relative error reduction of 56.36%, 20%, 4.5%, and 3.5% on the datasets of ICDAR-19, TableBank, UNLV, and Marmot, respectively. Furthermore, this paper sets a new benchmark by performing exhaustive cross-datasets evaluations to exhibit the generalization capabilities of the proposed method. [ABSTRACT FROM AUTHOR]

When measuring the electromagnetic parameters of flexible material with an air coaxial line, the specimen is prone to bend and deform in the fixture, which results in erroneous results. In order to solve this problem, this paper proposes a new measurement method. Firstly, a rigid material is selected and loaded into the air coaxial line for two-port S parameters measurement. Then, the flexible material is attached to one end of the rigid material and loaded into the air coaxial line together for repeated two-port S parameter measurement. According to the S parameters measured and the two-port network cascading theory, the S parameters of the flexible material separately loaded in the fixture can be deduced. Finally, the electromagnetic parameters of the flexible material can be calculated by the Nicolson−Ross−Weir (NRW) method. Experimental results show that the method proposed is effective and reliable. [ABSTRACT FROM AUTHOR]

Color constancy methods are generally based on a simplifying assumption that the spectral distribution of a light source is uniform across scenes. However, in reality, this assumption is often violated because of the presence of multiple light sources, that is, more than two illuminations. In this paper, we propose a unique cascade network of deep multi-scale supervision and single-scale estimation (CN-DMS4) to estimate multi-illumination. The network parameters are supervised and learned from coarse to fine in the training process and estimate only the final thinnest level illumination map in the illumination estimation process. Furthermore, to reduce the influence of the color channel on the Euclidean distance or the pixel-level angle error, a new loss function with a channel penalty term is designed to optimize the network parameters. Extensive experiments are conducted on single and multi-illumination benchmark datasets. In comparison with previous multi-illumination estimationmethods, our proposedmethod displays a partial improvement in terms of quantitative data and visual e􀀀ect, which provides the future research direction in end-to-end multi-illumination estimation. [ABSTRACT FROM AUTHOR]

High-frequency current transformers (HFCT) are widely used to measure fast transient current. Their advantages are simple structure and relatively moderate price. Their lower and upper bandwidth are limited, but the HFCT can be easily applied to many measuring applications in power electronics. The disadvantages of HFCT are substantial dimensions and a large weight. The paper proposes a system of a cascade connection of two transformers, which allows us to reduce these disadvantages. The properties of such an HFCT combination were investigated and described. In the article, the expression for double current transformer transmittances is derived. The frequency response of the sensor was determined, and the results were verified in a practical arrangement. An experimental setup of a cascade CT connection was made and tested, allowing for fast-changing signals in transients to be measured. This paper presents the theoretical basis and results of laboratory work on a wide range of static and dynamic tests of the proposed sensor. [ABSTRACT FROM AUTHOR]

Instance segmentation (IS) of remote sensing (RS) images can not only determine object location at the box-level but also provide instance masks at the pixel-level. It plays an important role in many fields, such as ocean monitoring, urban management, and resource planning. Compared with natural images, RS images usually pose many challenges, such as background clutter, significant changes in object size, and complex instance shapes. To this end, we propose a query-based RS image cascade IS network (QCIS-Net). The network mainly includes key components, such as the efficient feature extraction (EFE) module, multistage cascade task (MSCT) head, and joint loss function, which can characterize the location and visual information of instances in RS images through efficient queries. Among them, the EFE module combines global information from the Transformer architecture to solve the problem of long-term dependencies in visual space. The MSCT head uses a dynamic convolution kernel based on the query representation to focus on the region of interest, which facilitates the association between detection and segmentation tasks through a multistage structural design that benefits both tasks. The elaborately designed joint loss function and the use of the transfer-learning technique based on a well-known dataset (MS COCO) can guide the QCIS-Net in training and generating the final instance mask. Experimental results show that the well-designed components of the proposed method have a positive impact on the RS image instance segmentation task. It achieves mask average precision (AP) values of 75.2% and 73.3% on the SAR ship detection dataset (SSDD) and Northwestern Polytechnical University Very-High-Resolution dataset (NWPU-VHR-10 dataset), outperforming the other competitive models. The method proposed in this paper can enhance the practical application efficiency of RS images. [ABSTRACT FROM AUTHOR]

This paper is concerned with variational methods for open quantum systems with Markovian dynamics governed by Hudson–Parthasarathy quantum stochastic differential equations. These QSDEs are driven by quantum Wiener processes of the external bosonic fields and are specified by the system Hamiltonian and system–field coupling operators. We consider the system response to perturbations in these operators and introduce a transverse Hamiltonian which encodes the propagation of the perturbations through the unitary system–field evolution. This approach provides an infinitesimal perturbation analysis tool which can be used for the development of optimality conditions in quantum control and filtering problems. As performance criteria, such settings employ quadratic (or more complicated) cost functionals of the system and field variables to be minimized over the energy and coupling parameters of system interconnections. We demonstrate an application of the transverse Hamiltonian variational approach to a mean square optimal coherent quantum filtering problem for a measurement-free field-mediated cascade connection of a quantum system with a quantum observer. [ABSTRACT FROM AUTHOR]

A novel scheme for disambiguating conflicting classification results in Audio-Visual Speech Recognition applications is proposed in this paper. The classification scheme can be implemented with both generative and discriminative models and can be used with different input modalities, viz. only audio, only visual, and audio visual information. The proposed scheme consists of the cascade connection of a standard classifier, trained with instances of each particular class, followed by a complementary model which is trained with instances of all the remaining classes. The performance of the proposed recognition system is evaluated on three publicly available audio-visual datasets, and using a generative model, namely a Hidden Markov model, and three discriminative techniques, viz. random forests, support vector machines, and adaptive boosting. The experimental results are promising in the sense that for the three datasets, the different models, and the different input modalities, improvements in the recognition rates are achieved in comparison to other methods reported in the literature over the same datasets. [ABSTRACT FROM AUTHOR]

When studying an unfamiliar system, we first look for the symmetry that the system has so that we can make many predictions about the possible properties of the system. The symmetry in transport network security needs to maintain a stable state and maintain a constant state of transport network security. With the development of China–Europe freight trains, the transport between Asia and Europe has gradually formed the Belt and Road (B&R) land–sea transport network. In order to analyze the cascading failure mechanism of the B&R land–sea transport network, a network cascading failure model is constructed. Then, the quantitative analysis of the connectivity indicators of the land–sea transport network is conducted from the node attack strategy, and it is compared with the Maritime Silk Road (MSR) shipping network. Finally, the robustness of the land–sea transport network under emergencies is analyzed. From the results of deliberate attacks, the attack threshold of the B&R land–sea transport network is the same as that of the MSR shipping network, and the maximum number of attacks is slightly less than that of the MSR shipping network. The Russia–Ukraine conflict has a minimal impact on the robustness of cascading failure in the land–sea transport network. The Red Sea crisis may have a significant impact on the robustness of cascading failure in the land–sea transport network. The research results can provide suggestions for improving the robustness of the B&R land–sea transport network. [ABSTRACT FROM AUTHOR]

The networked combat system-of-system (CSOS) is the trend of combat development with the innovation of technology. To achieve the combat effectiveness, studying the ability of CSOS to cope with external interference is of great importance. Here we report a modeling method of CSOS from the perspective of complex networks and explore the robustness of the combat network based on this. Firstly, a more realistic double-layer heterogeneous dependent combat network model is established. Then, the conditional group dependency situation is considered to design failure rules for dependent failure, and the coupling relation between the double-layer subnets is analyzed for overload failure. Based on this, the initial load and capacity of the node are defined, respectively, as well as the load redistribution strategy and the status judgment rules for the cascading failure model. Simulation experiments are carried out by changing the attack modes and different parameters, and the results show that the robustness of the combat network can be effectively improved by improving the tolerance limit of one-way dependency of the functional net, the node capacity of the functional subnet, and the tolerance of the overload state. The conclusions of this paper can provide a useful reference for network structure optimization and network security protection in the military field. [ABSTRACT FROM AUTHOR]

This article proposes an approach and develops an appropriate method of applying linear non-equilibrium thermodynamics to analyze energy processes, in particular using the example of the wind energy conversion system (WECS) with a directly connected vertical axis wind turbine (VAWT) and vector-controlled permanent magnet synchronous generator (PMSG). The main steps of the proposed approach are the description of the component subsystems as universal linear or linearized energy converters (ECs), which are characterized by several dimensionless parameters, the main one of which is the degree of coupling between their input and output. According to their value, as well as justified efficiency criteria, the optimal operating points of each ECs can be easily found. Such an approach makes it possible to abstract from physical laws of a different nature and equally assess the work of each of the subsystems. The next step is a connection of the received ECs. As shown in the paper, for the most common cascade connection of ECs, there are the best conditions for their connection, under which the newly formed equivalent EC can have maximum efficiency. This opens up an opportunity to analyze the influence of already real parameters of cascaded interconnected subsystems on the quality of their connection and justify specific solutions that would not have been seen without this approach. For example, in this study, from all parameters of the PMSG, only the selection of the optimal rated inductance of the armature winding made it possible to improve the quality of the connection of the PMSG with a specific VAWT and approximate the efficiency of the entire WECS to the maximum possible, especially in medium and high winds. [ABSTRACT FROM AUTHOR]

With the continuous development of information technology, a spontaneous interdependent network has formed within the air traffic control network. Due to the internal interdependence, any small, failed node may trigger a cascade failure of the entire system. The purpose of this study is to investigate the resilience of air traffic control networks. Based on air traffic management regulations, a new cascading failure model for air traffic control networks is proposed, which is based on the theory of interdependent networks. The model establishes a dual-layer dependency relationship between the control coordination network and the air route facility network, including control dependency and service dependency. Through experiments, targeted measures are proposed to improve the safety and reliability of air traffic control. This model introduces parameters such as control cost and node control capability, and reflects the resilience of the air traffic control network, based on the final number of failed nodes after the steady-state of the cascade failure, the network's cascade failure rate, and the system's load failure threshold. Simulation results show that enhancing the control capability and increasing the number of control positions can improve the control cost of the air traffic control network. The higher the control cost, the better the resilience of the air traffic control network. Improving the control capability of control nodes has a greater impact on the resilience of the air traffic control network, compared to increasing the number of control nodes. The degree attack on route nodes has a greater impact on the cascade failure of the air traffic control network, compared to random attacks and facility node degree attacks. The cascade failure model proposed in this paper provides a new method for guiding the air traffic control network to resist cascade failure attacks and enhance its resilience. [ABSTRACT FROM AUTHOR]

The cascade connection of two NFSRs is an important class of NFSRs which has been used in the design of many recently proposed stream ciphers, such as Grain v1, Grain-128a, Fruit and Sprout. In this paper, we consider the decomposition of an NFSR into the cascade connection of two smaller NFSRs, which is equivalent to the factorization of a characteristic function h into f ∗ g where f and g are both characteristic functions. Inspired by the problem of factoring a large integer, we firstly consider a special case—that is, factoring h into the form h = g ∗ g . Based on the idea of "finding derivatives", we completely solve this case in theory, and present an algorithm to compute g from h. Moreover, if the characteristic function f is "close" to g, then one can also use the same algorithm to determine g and then get f by the existing method. For the general case, the idea of "finding derivatives" can also be used and some interesting results can be derived. [ABSTRACT FROM AUTHOR]

The key procedure is to accurately identify pedestrians in complex scenes and effectively embed features from multiple vision cues. However, it is still a limitation to coordinate two tasks in the unified framework, thus leading to high computational overhead and unsatisfactory search performance. Furthermore, most methods do not take significant clues and key features of pedestrians into consideration. To remedy these issues, we introduce a novel method named Multi-Attention-Guided Cascading Network (MGCN) in this paper. Specifically, we obtain the trusted bounding box through the detection header as the label information for post-process. Based on the end-to-end network, we demonstrate the advantages of jointly learning to construct the bounding box and attention module by maximizing the complementary information from different attention modules, which can achieve optimized person search performance. Meanwhile, by imposing an aligning module on re-id feature extracted network to locate visual clues with semantic information, which can restrain redundant background information. Extensive experimental results for the two benchmark person search datasets are provided to demonstrate that the proposed MGCN markedly outperforms the state-of-the-art baselines. [ABSTRACT FROM AUTHOR]

This paper aimed to increase accuracy of an Alzheimer's disease diagnosing function that was obtained in a previous study devoted to application of decision roots to the diagnosis of Alzheimer's disease. The obtained decision root is a discrete switching function of several variables applicated to aggregation of a few indicators to one integrated assessment presents as a superposition of few functions of two variables. Magnetic susceptibility values of the basal veins and veins of the thalamus were used as indicators. Two categories of patients were used as function values. To increase accuracy, the idea of using artificial neural networks was suggested, but a feature of medical data is its limitation. Therefore, neural networks based on limited training datasets may be inefficient. The solution to this problem is proposed to preprocess initial datasets to determine the parameters of the neural networks based on decisions' roots, because it is known that any can be represented in the incompletely connected neural network form with a cascade structure. There are no publicly available specialized software products allowing the user to set the complex structure of a neural network, which is why the number of synaptic coefficients of an incompletely connected neural network has been determined. This made it possible to predefine fully connected neural networks, comparable in terms of the number of unknown parameters. Acceptable accuracy was obtained in cases of one-layer and two-layer fully connected neural networks trained on limited training sets on an example of diagnosing Alzheimer's disease. Thus, the scientific hypothesis on preprocessing initial datasets and neural network architecture selection using special methods and algorithms was confirmed. [ABSTRACT FROM AUTHOR]

With the evolution of the Internet and information technology, the era of big data is a new digital one. Accordingly, animation IP has been more and more widely welcomed and concerned with the continuous development of the domestic and international animation industry. Hence, animation video analysis will be a good landing application for computers. This paper proposes an algorithm based on clustering and cascaded SSD for object detection of animation characters in the big data environment. In the training process, the improved classification Loss function based on Focal Loss and Truncated Gradient was used to enhance the initial detection effect. In the detection phase, this algorithm designs a small target enhanced detection module cascaded with an SSD network. In this way, the high-level features corresponding to the small target region can be extracted separately to detect small targets, which can effectively enhance the detection effect of small targets. In order to further improve the effect of small target detection, the regional candidate box is reconstructed by a k-means clustering algorithm to improve the detection accuracy of the algorithm. Experimental results demonstrate that this method can effectively detect animation characters, and performance indicators are better than other existing algorithms. [ABSTRACT FROM AUTHOR]

The practical application of wireless sensor networks (WSNs) in hot fields is summarized. It is found that compared with traditional monitoring methods, it has better adaptability to complex environments and low cost. It is suitable for monitoring power grid operation parameters. Therefore, this paper combines the above network and cascading failures, analyzes its 24-hour continuous and dynamic monitoring of the operation parameters of the power grid, and considers how to use the obtained parameters to analyze the disturbance of the remaining lines after the initial fault of the power grid. To prevent cascading failures in the power grid, a preventive control model considering safety and economy is proposed, and the model is solved by nondominated sorting genetic algorithm II and particle swarm optimization (NSGA2-PSO). Finally, the rationality of this method is verified in the IEEE39 node system. [ABSTRACT FROM AUTHOR]

This paper focuses on long-tailed object detection in the semi-supervised learning setting, which poses realistic challenges, but has rarely been studied in the literature. We propose a novel pseudo-labeling-based detector called CascadeMatch. Our detector features a cascade network architecture, which has multi-stage detection heads with progressive confidence thresholds. To avoid manually tuning the thresholds, we design a new adaptive pseudo-label mining mechanism to automatically identify suitable values from data. To mitigate confirmation bias, where a model is negatively reinforced by incorrect pseudo-labels produced by itself, each detection head is trained by the ensemble pseudo-labels of all detection heads. Experiments on two long-tailed datasets, i.e., LVIS and COCO-LT, demonstrate that CascadeMatch surpasses existing state-of-the-art semi-supervised approaches—across a wide range of detection architectures—in handling long-tailed object detection. For instance, CascadeMatch outperforms Unbiased Teacher by 1.9 AP Fix on LVIS when using a ResNet50-based Cascade R-CNN structure, and by 1.7 AP Fix when using Sparse R-CNN with a Transformer encoder. We also show that CascadeMatch can even handle the challenging sparsely annotated object detection problem. Code: https://github.com/yuhangzang/CascadeMatch. [ABSTRACT FROM AUTHOR]

The highlighted area or object is defined as the salient region or salient object. For salient object detection, the main challenges are still the clarity of the boundary information of the salient object and the positioning accuracy of the salient object in the complex background, such as noise and occlusion. To remedy these issues, it is proposed that the asynchronous cascade saliency detection algorithm based on a deep network, which is embedded in an encoder–decoder architecture. Moreover, the lightweight hybrid attention module is designed to obtain the explicit boundaries of salient regions. In order to effectively improve location information of salient objects, this paper adopts a bi-directional asynchronous cascade fusion strategy, which generates prediction maps with higher accuracy. The experimental results on five benchmark datasets show that the proposed network HACNet is on a par with the state of the art for image saliency datasets. [ABSTRACT FROM AUTHOR]

Deep-learning-based registration methods can not only save time but also automatically extract deep features from images. In order to obtain better registration performance, many scholars use cascade networks to realize a coarse-to-fine registration progress. However, such cascade networks will increase network parameters by an n-times multiplication factor and entail long training and testing stages. In this paper, we only use a cascade network in the training stage. Unlike others, the role of the second network is to improve the registration performance of the first network and function as an augmented regularization term in the whole process. In the training stage, the mean squared error loss function between the dense deformation field (DDF) with which the second network has been trained and the zero field is added to constrain the learned DDF such that it tends to 0 at each position and to compel the first network to conceive of a better deformation field and improve the network's registration performance. In the testing stage, only the first network is used to estimate a better DDF; the second network is not used again. The advantages of this kind of design are reflected in two aspects: (1) it retains the good registration performance of the cascade network; (2) it retains the time efficiency of the single network in the testing stage. The experimental results show that the proposed method effectively improves the network's registration performance compared to other state-of-the-art methods. [ABSTRACT FROM AUTHOR]

Dynamic facial expression recognition (DFER) is a promising research area because it concerns the dynamic change pattern of facial expressions, but it is difficult to effectively capture the facial appearances and dynamic temporal information of each image in an image sequence. In this paper, a cascaded spatiotemporal attention network (CSTAN) is proposed to learn and integrate spatial and temporal emotional information in the process of facial expression change. Three types of attention modules are embedded into the cascaded network to enable it to extract more informative spatiotemporal features for the DFER task in different dimensions. A channel attention module helps the network focus on the meaningful spatial feature maps for the DFER task, a spatial attention module focuses on the regions of interest among the spatial feature maps, and a temporal attention module aims to explore the dynamic temporal information when an expression changes. The experimental results on three public facial expression recognition datasets prove the good performance of the CSTAN, and it can extract representative spatiotemporal features. Meanwhile, the visualization results reveal that the CSTAN can locate regions of interest and contributing timesteps, which illustrates the effectiveness of the multidimensional attention modules. [ABSTRACT FROM AUTHOR]