Search

Intuitively, global video context could benefit video semantic segmentation (VSS) if it is designed to simultaneously model global temporal and spatial dependencies for a holistic understanding of the semantic scenes in a video clip. However, we found that the existing VSS approaches focus only on modeling local video context. This paper attempts to bridge this gap by learning global video context for VSS. Apart from the global nature, the video context should also be compact when considering the large number of video feature tokens and the redundancy among nearby video frames. Then, we embed the learned global and compact video context into the features of the target video frame to improve the distinguishability. The proposed VSS method is dubbed Global and Compact Video Context Embedding (GCVCE). With the compact nature, the number of global context tokens is very limited so that GCVCE is flexible and efficient for VSS. Since it may be too challenging to directly abstract a large number of video feature tokens into a small number of global context tokens, we further design a Cascaded Convolutional Downsampling (CCD) module before GCVCE to help it work better. 1.6% improvement in mIoU on the popular VSPW dataset compared to previous state-of-the-art methods demonstrate the effectiveness and efficiency of GCVCE and CCD for VSS. Code and models will be made publicly available.

Hyperspectral images (HSIs) are prone to be corrupted by various types of noise during the process of imaging and transmission, which seriously affect the subsequent HSI processing tasks. In this article, we proposed a novel low-rank-based model for HSIs denoising. On one hand, motivated by the superiority of nonconvex approximation to matrix rank, we construct a new nonconvex function which is tighter than some of existing rank approximated functions. On the other hand, we describe Gaussian noise and sparse noise simultaneously by introducing correntropy. In comparison with traditional model, we constrain noise in one regularization instead of separately constraining Gaussian noise and sparse noise, which indicates that the number of regularization parameters have been reduced. To optimize the proposed model, some convex analysis tools are utilized in this article. In addition, we provide theoretical analysis on the convergence of the developed algorithm. Through experiments conducted on both simulated and real HSIs, we verify the superiority of our model in enhancing the performance of mixture noise removal in HSIs.

Motion segmentation is a formidable computer vision task, aiming to segment moving targets from a dynamic scene. In this paper, we choose to introduce an additional modality to bolster the robustness. The event camera is a bio-inspired sensor that accurately detects and captures intensity changes with exceptional temporal resolution and dynamic range, which is an optimal choice for motion segmentation. Therefore, we present a novel framework for event-based motion segmentation and propose Multi-Scale Recurrent Neural Network (MSRNN) to fuse temporal information efficiently. To our best knowledge, it is the first time that a multi-scale recurrent architecture is implemented in event-based motion segmentation. The proposed framework is evaluated through experiments conducted on the EV-IMO dataset. Our method achieves a mean Intersection-over-Union (mIoU) of 82.0%, which sets a new state-of-the-art in motion segmentation. To further validate our approach in arduous real-world scenarios, we introduce the Event Challenging Motion dataset, consisting of 350 images and corresponding events, in which our method outperforms the other methods by 1.5% in Intersection-over-Union (IoU).

Epilepsy as a common disease of the nervous system, with high incidence, sudden and recurrent characteristics. Therefore, timely prediction of seizures and intervention treatment can significantly reduce the accidental injury of patients and protect the life and health of patients. Epilepsy seizures is the result of temporal and spatial evolution, Existing deep learning methods often ignore its spatial features, in order to make better use of the temporal and spatial characteristics of epileptic EEG signals. We propose a CBAM-3D CNN-LSTM model to predict epilepsy seizures. First, we apply short-time Fourier transform(STFT) to preprocess EEG signals. Secondly, the 3D CNN model was used to extract the features of preictal stage and interictal stage from the preprocessed signals. Thirdly, Bi-LSTM is connected to 3D CNN for classification. Finally CBAM is introduced into the model. Different attention is given to the data channel and space to extract key information, so that the model can accurately extract interictal and pre-ictal features. Our proposed approach achieved an accuracy of 97.95%, a sensitivity of 98.40%, and a false alarm rate of 0.017 h−1 on 11 patients from the public CHB-MIT scalp EEG dataset. Clinical and Translational Impact Statement—Timely prediction of epileptic seizures and intervention treatment can significantly reduce the accidental injury of patients and protect the life and health of patients.

Due to the complexity of terrain in natural environments, the soft exoskeleton cannot adaptively adjust parameters to achieve the optimal performance. To this end, a design for a soft exoskeleton assistive force parameter optimization method on multi-terrains is presented in this paper. Firstly, the core control parameters are determined by analyzing the system’s motion dynamics. Then, the collected data from inertial measurement unit (IMU) is transferred to the convolutional neural network (CNN) to recognize the certain terrain. In the meanwhile, the control parameters corresponding to the different terrains are optimized by the Bayesian algorithm. Finally, the optimal assistive force parameters are transferred to the system for improving the performance of the soft exoskeleton. The experiment is conducted on three participants, wherein the net metabolic rates of the subjects are compared with and without the assistive force. The final results show that the metabolic rates of the subjects reduce the average value of 19.6% on flat ground, 11.6% on walking uphill, and 12.7% on walking upstairs. The experimental results confirm the effectiveness of the proposed method.

3D reconstruction is the transformation of real objects into mathematical models. By using 3D models, we can observe the shape and measure the parameters, and help us to analyze the properties of objects. For the problems of incompleteness and inefficiency in the reconstruction of object 3D point clouds, a fast and automated system for panoramic 3D point cloud reconstruction of objects was proposed. First, we designed an automatic platform, which could acquire RGB image sequences of objects in two directions. Then we adopted the Structure From Motion (SFM) algorithm to generate point clouds. For the problem of different scales of point clouds, we obtained the scaling by calculating the length ratio of the axes of the oriented bounding box, and scaled the point clouds to a uniform scale. In addition, markers were placed around the object and used to acquire the rotation matrix of the object point cloud in two directions. Finally, we verified the point cloud models of different objects generated by the system, and found that the relative error didn’t exceed 6.67%. According to the results, the system proposed could reconstruct the panoramic 3D point cloud of the object better and provide a reference for related research.

Standalone solar PV systems have emerged as potential alternatives to electricity problems in areas where a grid is unavailable. Obtaining full power from a photoelectric system, DC-DC inverter, DC-AC converter, and control system presents great difficulties when building these devices. A standalone two-stage approach is introduced in this work with a boost converter followed by an inverter and a battery with a bidirectional converter. In this paper, a novel virtual synchronous generator (VSG) controller is designed and implemented to adjust the inverter output. The VSG element and the maximum power point tracking (MPPT) used in this study serve the following purposes: to adjust the inverter output and to realize the maximum power of the PV scheme. The new control strategy design was evaluated and validated using extensive MATLAB simulations under different scenarios, including load variations. The system output was evaluated using extensive MATLAB simulations. A hands-on experiment was conducted for the VSG using console testing. Due to the lack of laboratory equipment, we could not experiment with the entire system.

3D point cloud registration ranks among the most fundamental problems in remote sensing, photogrammetry, robotics and geometric computer vision. Due to the limited accuracy of 3D feature matching techniques, outliers may exist, sometimes even in very large numbers, among the correspondences. Since existing robust solvers may encounter high computational cost or restricted robustness, we propose a novel, fast and highly robust solution, named VOCRA (VOting with Cost function and Rotating Averaging), for the point cloud registration problem with extreme outlier rates. Our first contribution is to employ the Tukey’s Biweight robust cost to introduce a new voting and correspondence sorting technique, which proves to be rather effective in distinguishing true inliers from outliers even with extreme (99%) outlier rates. Our second contribution consists in designing a time-efficient consensus maximization paradigm based on robust rotation averaging, serving to seek inlier candidates among the correspondences. Finally, we apply Graduated Non-Convexity with Tukey’s Biweight (GNC-TB) to estimate the correct transformation with the inlier candidates obtained, which is then used to find the complete inlier set. Both standard benchmarking and realistic experiments with application to two real-data problems are conducted, and we show that our solver VOCRA is robust against over 99% outliers (exceeding traditional GNC by nearly 10% and RANSAC by nearly 4%) and more time-efficient than the state-of-the-art competitors.

When users upload their private data to the cloud, they lose control of the data stored in the cloud server. If the cloud system cannot provide an effective security mechanism to protect the data, the consequent data leakage issue will hinder the development of cloud computing. Conventional access control and encryption technologies cannot effectively control the propagation of tenant private data in the system. The mandatory one-way information flow control model is limited by the complexity of the cloud environment, and it is difficult to effectively protect private data stored in the cloud. To solve the above problems, this article proposes a tenant-led ciphertext information flow control method for cloud virtual machines. Through the design of a decentralized information flow control security policy, a secret-domain key management scheme, and a multi-ID-based threshold encryption scheme, the information flow control strategies of taint infection, secret-level reduction, and ability propagation are realized in a ciphertext form, which can effectively prevent malicious users inside and outside the system from illegally reading private data. The feasibility of this method is verified by a security proof and an experiment.

Field-programmable photonic array-based chips will be key components in realizing high-performance THz communication in the future. However, the optical diffraction limit prohibits their integration in chip-level sizes. Although plasmonic waveguides combined with graphene are able to possess mode transmission at THz wavelengths, their operational problems such as complex device structures or high transmission losses still cannot be addressed effectively. In this paper, by relying on a mechanism of THz surface plasmon mode of graphene and an energy distribution mechanism of the refractive index difference region, a novel hybrid graphene plasmonic waveguide is proposed, whose core design part is a simple gap-slot region. This waveguide's transmission distance is one order of magnitude longer than that of the conventional waveguide while retaining the benefit high energy confinement, under appropriate parameters. Hence the proposed waveguide can be an efficient candidate component of field-programmable photonic arrays for THz communication.

Rotation symmetric Boolean functions (RSBFs) are used widely in symmetric cryptography. In this paper, a systematic construction of balanced odd-variable RSBFs satisfying strict avalanche criterion is proposed. Hence, a class of $(6k+3)$ -variable resilient RSBFs satisfying strict avalanche criterion is also presented for any $k\geq 2$ . Some of the obtained RSBFs have many other good cryptographic properties at the same time, that is, optimal algebraic degree, good global avalanche characteristics, high nonlinearity and nonexistence of nonzero linear structures. Moreover, we obtain some count results of RSBFs satisfying strict avalanche criterion. This is the first time that the autocorrelation properties of RSBFs are investigated systemically.

Previous deep learning based approaches to text baseline detection in historical documents usually take it as a semantic segmentation task. These methods adopt a fully convolutional neural network to predict baseline pixels first and then group them into lines by heuristic post-processing steps, which tends to suffer from a wrongly merged or wrongly split problem owing to limited context information provided by pixels. To address these issues, we introduce the concept of a baseline primitive, which is defined as a virtual bounding box centered at each baseline pixel. After baseline primitive detection, a relation network is used to predict a link relationship for each pair of primitives. Consequently, text baselines are generated by detecting baseline primitives and grouping them with the corresponding link relationships. Owing to the design of baseline primitives, wider context information can be leveraged to improve link prediction accuracy. Therefore, our approach can effectively detect text baselines with small inter-line or large inter-word spacing. Quantitative experimental results demonstrate the effectiveness of the proposed baseline primitive design. Our approach achieves state-of-the-art performance on two public benchmarks, namely cBAD 2017 and cBAD 2019.

Hill musculoskeletal model (HMM) is commonly used to estimate human motion intentions. HMM utilizes electromyography (EMG) signals as the nonlinear model input to obtain muscle forces or torques. However, due to the fact that it contains many physiological parameters that are difficult to measure, HMM is generally applied in simple continuous intention estimation of a single joint. In this work, we aimed at recognizing shoulder and elbow joints angles and their angular velocities continuously in real time. Firstly, we used MYO armband as the EMG sensor. Then, a reasonable prediction model was deduced based on HMM and human dynamics to realize online continuous recognition of the four angles and angular velocities of shoulder and elbow joints. Nonlinear autoregressive with external input neural network (NARX) replaced the prediction equation. In addition, the framework of state space model was completed by constructing an observation equation. Thus, the closed-loop characteristic was realized to eliminate the influence of cumulative error and ensure good estimation performance. Experimental results verified the feasibility and accuracy of the algorithm. For predefined trajectory and random trajectory separately, the RMSE were 0.955 and 1.15 (degree) for angles estimation and 2.8, 3.40 for angular velocities (degree/s). Compared with the normally used back-propagation neural network (BPNN), the method proposed in this paper obviously got more accurate and smooth results.

The Perspective-n-Point-and-Line (PnPL) problem, as a cornerstone in geometric computer vision, seeks to estimate the absolute pose of a calibrated camera from 3D-to-2D point and line correspondences. In this paper, we present a certifiably globally optimal and robust solution to the PnPL problem with a large number of outliers among the correspondences. Our first contribution is to reformulate the general PnPL problem as a novel constrained global optimization problem of Sum-of-Squares (SOS) polynomials using generalized geometric distances for both points and lines. Our second contribution is to efficiently solve this non-convex optimization problem by reducing it to an equivalent convex Linear Matrix Inequality (LMI) problem via a series of SOS relaxations. With these two contributions, we can develop a non-minimal solver, named SPnPL, for the outlier-free cases. The third contribution is to further add a Graduated Non-Convexity (GNC) cost function to SPnPL so as to remove outliers through closed-form iterations, which leads to a robust solver, named GNC-SPnPL. Both synthetic and real-data experiments confirm that SPnPL can mostly outperform the existing state-of-the-art PnPL algorithms and GNC-SPnPL can render accurate results even when 85% of the correspondences are outliers.

Three-phase voltage-source pulse width modulation (PWM) rectifier has been widely employed in various industrial applications. For the traditional dual-loop PI control of three-phase voltage-source PWM rectifier, the voltage loop and current loop have poor disturbance rejection performance and long response time. A linear active disturbance rejection control (LADRC) strategy is proposed for three-phase voltage-source PWM rectifier in this paper. The coupling variable and external disturbance in the model of PWM rectifier are observed and compensated by linear extended state observer (LESO) in two-phase synchronous reference frame. The LADRC outer-voltage loop controller and LADRC inner-current loop controller of PWM rectifier are designed. The proposed control strategy is compared and verified with PI control and ADRC by MATLAB/Simulink. Simulation and experimental results confirm the superiority of proposed control strategy in dynamic performance and disturbance rejection.

Aiming at the problem of fault prognostics for the energy storage power station, this paper proposes a novel data-driven method named multiple elastic networks with time delays (MEN-TD). The proposed method can learn the status of the energy storage power station in advance and provide early detection of the fault. First, through the correlation analysis and the mechanism knowledge, the energy storage power station key parameter and corresponding key factors affecting the parameter are determined. Secondly, in order to predict the trend of the key parameter over a period of time and improve the prediction accuracy, the MEN-TD model is constructed. Then, based on the predicted values of the key parameter, compared with the control limit in the healthy status, the fault can be pre-warned in advance. Finally, through testing on the practical energy storage power station in Zhenjiang of China, the effectiveness and superiority of the proposed MEN-TD method are demonstrated.

In this paper, the semi-global cooperative output regulation problem of heterogeneous swarm systems subject to saturation under switching network is investigated. Like the common consensus problems, the swarm systems consist of some followers cannot access the state of the leader directly. But the difference is that some even cannot measure the state of the leader all the time. By utilizing the distributed output feedback control method, the low-gain feedback technique and the output regulation theories, we establish two sets of consensus control laws, respectively, based on the state feedback method and the output feedback method, to solve the output consensus problems in this kind. Besides, the global leader-following output consensus problem under switching network is proved to be a special case of semi-global cooperative output regulation problem under switching network, and as a consequence, it can also be solved by our method. A simulation example is proposed to illustrate the results very well.

Increasing deployment of distributed energy resources (DERs) is re-sculpturing the modern power systems in recent years. Future smart power distribution systems should be competent at accommodating extensive integration of DERs and managing the associated uncertainties at the distribution level. The electricity market has been proved to be an efficient way to employ market signals to direct behaviors of users and DERs with large capacity and homogeneous pattern. However, existing market frameworks cannot effectively handle a large number of small-scale DERs due to their diverse characteristics and arbitrary behavior patterns. In this context, an aggregated model which can represent and manage a diverse collection of DER, load, and storage is proposed. An additional trading platform, namely the energy sharing market, is established to reinforce the coordination and collaboration among various aggregators as well as operators. Energy sharing scheme is applied and a corresponding dynamic dispatch platform is designed to solve the crowdsource problem. The efficiency of the proposed model is validated by the numerical studies, and the market performance and impacts of energy sharing on the power systems are illustrated.

In this paper, we propose a unified array geometry, dubbed generalized nested subarray (GNSA), for the underdetermined direction-of-arrival estimation. The GNSA is composed of multiple, identical subarrays, which can be a minimum redundancy array (MRA), a (super) nested array, a uniform linear array (ULA), or any other linear arrays with hole-free difference coarrays (DCAs). By properly design the spacings between subarrays, the resulting DCA of the GNSA can also be a hole-free (filled) ULA. When the subarray is an MRA and meanwhile its sensors' positions also follow an MRA configuration, a nested MRA (NMRA) is constructed. This NMRA can provide up to O(M2N2) degrees of freedom (DOFs) using only MN physical sensors. In order to fully utilize the increased DOF, we develop a new DOA estimation algorithm, which consists of a dimensional reduction matrix to exploit the data of all virtual elements, a Toeplitz matrix to decorrelate the equivalent coherent sources, and a root-MUSIC method to mitigate the computational workload. This new algorithm can achieve better DOA estimation performance than traditional spatial smoothing MUSIC algorithm with lower computational complexity. Numerical simulation results demonstrate the superiorities of the proposed array geometry in resolving more sources than sensors, DOA estimation performance, and the angular resolution.

In recent years, role-based access control (RBAC) has become the de facto access control model due to its good applicability and high flexibility. Since the organizations need to update the access control policies to meet the changes in employees, departments, business processes, and so on. The RBAC system has to define new roles and becomes more and more bloated because it's difficult to modify the role-permission assignment with no or minimal impact to other users and roles. Hence, there is a great need to reconfigure the RBAC system over time to reduce its structural complexity and keep as close as possible to the original. Several RBAC reconfiguration approaches have been proposed aiming at generating roles similar to the deployed ones, but they neglect the differences in deployed roles that some of them are useless for the system and generate more roles than needed, which in turn increases the system structure complexity. In this paper, we first propose three indicators to evaluate the quality of deployed roles and define the problem of hierarchy RBAC reconfiguration with minimal weight structure complexity and perturbation. Then, the hierarchy RBAC reconfiguration approach and its algorithm process are proposed to address the problem. To conclude, we demonstrate the effectiveness and stability of our approach through experiments.

In this paper, we propose a microdisk resonator with negative thermal optical coefficient (TOC) polymer for refractive index (RI) sensing with thermal stability. The transmission characteristics and sensing performances by using quasi-TE01 and quasi-TM01 modes are simulated by a three-dimensional finite element method. The influences of the TOC, RI, and thickness of the polymer on the sensing performances are also investigated. The simulation results show that the RI sensitivity Sn and temperature sensitivity ST with different polymers are in the ranges of 25.1-26 nm/RIU and 67.3-75.2 pm/K for the quasi-TE01 mode, and 94.5-110.6 nm/RIU and 1.2-51.3 pm/K for the quasi-TM01 mode, respectively. Moreover, figure-of-merit of the temperature sensing for the quasi-TM01 mode is in the range of 2 × 10-4-8 × 10-3, which can find important application in the implementation of the adiabatic devices.

In this paper, a microdisk resonator with two whispering-gallery modes (WGMs) is proposed for label-free biochemical sensing. According to the transmission responses of the two WGMs with different coupling gaps, there is the critical coupling status for the WGM (0, 36) while there is no critical coupling status for the WGM (1, 28). For the WGMs (0, 36) and (1, 28), the refractive index (RI) sensitivities of 45.8821 and 72.9402 nm/RIU are obtained, and the corresponding RI detection limits (DLs) of 8.5902 × 10-4 and 1.9228 × 10-3 RIU are achieved, respectively. Moreover, the proposed sensor also has the temperature sensitivities of 0.0730 and 0.0703 nm/K, which correspond to the temperature DLs of 0.1631 and 0.6263 K, respectively. By constructing a characteristic matrix, it is demonstrated that simultaneous measurement of the RI and temperature can be achieved.

The humidity and temperature responses of a microfiber coupler (MFC) coated with silica gel are investigated. Two MFC structures with different waist diameters of 2.5 and 3.5 μm were fabricated by fusing and tapering two single-mode fibers using a microheater brushing technique. The influences of the coating thickness and tapered waist diameter on the sensing performance are analyzed. For the proposed sensor with a waist diameter of 2.5 μm and 8-layers thick coating, the change in the relative humidity (RH) results in an exponential blueshift with a maximum sensitivity of 1.6 nm/% RH in the range from 70 to 86% RH. In response to the temperature change, the sensor's transmission spectrum redshifts in a linear fashion with an average sensitivity of 0.55 nm/°C in the range from 20 to 40 °C. The study is important for the development of the proposed fiber structure as a humidity or temperature sensor.