Unmanned aerial vehicles(UAVs) have gained significant attention in practical applications, especially the low-altitude aerial(LAA) object detection imposes stringent requirements on recognition accuracy and computati...
详细信息
Unmanned aerial vehicles(UAVs) have gained significant attention in practical applications, especially the low-altitude aerial(LAA) object detection imposes stringent requirements on recognition accuracy and computational resources. In this paper, the LAA images-oriented tensor decomposition and knowledge distillation-based network(TDKD-Net) is proposed,where the TT-format TD(tensor decomposition) and equalweighted response-based KD(knowledge distillation) methods are designed to minimize redundant parameters while ensuring comparable performance. Moreover, some robust network structures are developed, including the small object detection head and the dual-domain attention mechanism, which enable the model to leverage the learned knowledge from small-scale targets and selectively focus on salient features. Considering the imbalance of bounding box regression samples and the inaccuracy of regression geometric factors, the focal and efficient IoU(intersection of union) loss with optimal transport assignment(F-EIoU-OTA)mechanism is proposed to improve the detection accuracy. The proposed TDKD-Net is comprehensively evaluated through extensive experiments, and the results have demonstrated the effectiveness and superiority of the developed methods in comparison to other advanced detection algorithms, which also present high generalization and strong robustness. As a resource-efficient precise network, the complex detection of small and occluded LAA objects is also well addressed by TDKD-Net, which provides useful insights on handling imbalanced issues and realizing domain adaptation.
Wide-bandgap semiconductors exhibit much larger energybandgaps than traditional semiconductors such as silicon,rendering them very promising to be applied in the fields of electronics and *** examples of semiconductor...
详细信息
Wide-bandgap semiconductors exhibit much larger energybandgaps than traditional semiconductors such as silicon,rendering them very promising to be applied in the fields of electronics and *** examples of semiconductors include SiC,GaN,ZnO,and diamond,which exhibitdistinctive characteristics such as elevated mobility and *** characteristics facilitate the operation of awide range of devices,including energy-efficient bipolar junctiontransistors(BJTs)and metal-oxide-semiconductor field-effecttransistors(MOSFETs),as well as high-frequency high-electronmobility transistors(HEMTs)and optoelectronic components suchas light-emitting diodes(LEDs)and *** semiconductorsare used in building integrated circuits(ICs)to facilitate theoperation of power electronics,computer devices,RF systems,andother optoelectronic *** breakthroughs includevarious applications such as imaging,optical communication,*** them,the field of power electronics has witnessedtremendous progress in recent years with the development of widebandgap(WBG)semiconductor devices,which is capable ofswitching large currents and voltages rapidly with low ***,it has been proven challenging to integrate these deviceswith silicon complementary metal oxide semiconductor(CMOS)logic circuits required for complex control *** monolithic integration of silicon CMOS with WBG devices increases thecomplexity of fabricating monolithically integrated smart integrated circuits(ICs).This review article proposes implementingCMOS logic directly on the WBG platform as a ***,achieving the CMOS functionalities with the adoption of WBGmaterials still remains a significant *** article summarizesthe research progress in the fabrication of integrated circuitsadopting various WBG materials ranging from SiC to diamond,with the goal of building future smart power ICs.
Heterogeneous and monolithic integration of the versatile low-loss silicon nitride platform with low-temperature materials such as silicon electronics and photonics,III–V compound semiconductors,lithium niobate,organ...
详细信息
Heterogeneous and monolithic integration of the versatile low-loss silicon nitride platform with low-temperature materials such as silicon electronics and photonics,III–V compound semiconductors,lithium niobate,organics,and glasses has been inhibited by the need for high-temperature annealing as well as the need for different process flows for thin and thick *** techniques are needed to maintain the state-of-the-art losses,nonlinear properties,and CMOS-compatible processes while enabling this next generation of 3D silicon nitride *** report a significant advance in silicon nitride integrated photonics,demonstrating the lowest losses to date for an anneal-free process at a maximum temperature 250℃,with the same deuterated silane based fabrication flow,for nitride and oxide,for an order of magnitude range in nitride thickness without requiring stress mitigation or *** report record low anneal-free losses for both nitride core and oxide cladding,enabling 1.77 dBm^(-1) loss and 14.9 million Q for 80 nm nitride core waveguides,more than half an order magnitude lower loss than previously reported sub 300℃ *** 800 nm-thick nitride,we achieve as good as 8.66 dBm^(-1) loss and 4.03 million Q,the highest reported Q for a low temperature processed resonator with equivalent device area,with a median of loss and Q of 13.9 dBm^(-1) and 2.59 million each *** demonstrate laser stabilization with over 4 orders of magnitude frequency noise reduction using a thin nitride reference cavity,and using a thick nitride micro-resonator we demonstrate OPO,over two octave supercontinuum generation,and four-wave mixing and parametric gain with the lowest reported optical parametric oscillation threshold per unit resonator *** results represent a significant step towards a uniform ultra-low loss silicon nitride homogeneous and heterogeneous platform for both thin and thick waveguides capable of linear and nonlinear photonic circuits and
Utilization of FPGAs has increased dramatically in various application domains, mainly due to their unique traits. Especially, FPGAs' dynamic partial reconfiguration feature enables multiple and complex/large appl...
详细信息
Fog computing is a key enabling technology of 6G systems as it provides quick and reliable computing,and data storage services which are required for several 6G *** Intelligence(AI)algorithms will be an integral part ...
详细信息
Fog computing is a key enabling technology of 6G systems as it provides quick and reliable computing,and data storage services which are required for several 6G *** Intelligence(AI)algorithms will be an integral part of 6G systems and efficient task offloading techniques using fog computing will improve their performance and *** this paper,the focus is on the scenario of Partial Offloading of a Task to Multiple Helpers(POMH)in which larger tasks are divided into smaller subtasks and processed in parallel,hence expediting task ***,using POMH presents challenges such as breaking tasks into subtasks and scaling these subtasks based on many interdependent factors to ensure that all subtasks of a task finish simultaneously,preventing resource ***,applying matching theory to POMH scenarios results in dynamic preference profiles of helping devices due to changing subtask sizes,resulting in a difficult-to-solve,externalities *** paper introduces a novel many-to-one matching-based algorithm,designed to address the externalities problem and optimize resource allocation within POMH ***,we propose a new time-efficient preference profiling technique that further enhances time optimization in POMH *** performance of the proposed technique is thoroughly evaluated in comparison to alternate baseline schemes,revealing many advantages of the proposed *** simulation findings indisputably show that the proposed matching-based offloading technique outperforms existing methodologies in the literature,yielding a remarkable 52 reduction in task latency,particularly under high workloads.
This paper presents a controller for fast and ultrafast electric vehicle(EV)charging *** affecting the charging efficiency,the proposed controller enables the charger to provide support to the interconnection voltage ...
详细信息
This paper presents a controller for fast and ultrafast electric vehicle(EV)charging *** affecting the charging efficiency,the proposed controller enables the charger to provide support to the interconnection voltage to counter and damp its *** solutions are either hardware-based such as using supercapacitors and flywheels which increase the cost and bulkiness of the charging station,or software-based such as P/V droop methods which are still unable to provide a robust and strong voltage *** paper proposes an emulated supercapacitor concept in the control system of the ultra-fast EV charger in an islanded DC ***,it converts the EV from a static load to a bus voltage supportive load,leading to reduced bus voltage oscillations during single and multiple ultra-fast EV charging operations,and rides through and provides supports during extreme external *** analysis and design guidelines of the proposed controller are presented,and its effectiveness and improved performance compared with conventional techniques are shown for different case studies.
Quantum power system state estimation(QPSSE)offers an inspiring direction for tackling the challenge of state estimation through quantum ***,the current bottlenecks originate from the scarcity of practical and scalabl...
详细信息
Quantum power system state estimation(QPSSE)offers an inspiring direction for tackling the challenge of state estimation through quantum ***,the current bottlenecks originate from the scarcity of practical and scalable QPSSE methodologies in the noisy intermediate-scale quantum(NISQ)*** paper devises a NISQ−QPSSE algorithm that facilitates state estimation on real NISQ *** new contributions include:(1)A variational quantum circuit(VQC)-based QPSSE formulation that empowers QPSSE analysis utilizing shallow-depth quantum circuits;(2)A variational quantum linear solver(VQLS)-based QPSSE solver integrating QPSSE iterations with VQC optimization;(3)An advanced NISQ-compatible QPSSE methodology for tackling the measurement and coefficient matrix issues on real quantum computers;(4)A noise-resilient method to alleviate the detrimental effects of noise *** encouraging test results on the simulator and real-scale systems affirm the precision,universal-ity,and scalability of our QPSSE algorithm and demonstrate the vast potential of QPSSE in the thriving NISQ era.
Estimating the crowd count and density of highly dense scenes witnessed in Muslim gatherings at religious sites in Makkah and Madinah is critical for developing control strategies and organizing such a large ***,since...
详细信息
Estimating the crowd count and density of highly dense scenes witnessed in Muslim gatherings at religious sites in Makkah and Madinah is critical for developing control strategies and organizing such a large ***,since the crowd images in this case can range from low density to high density,detection-based approaches are hard to apply for crowd ***,deep learning-based regression has become the prominent approach for crowd counting problems,where a density-map is estimated,and its integral is further computed to acquire the final count *** this paper,we put forward a novel multi-scale network(named 2U-Net)for crowd counting in sparse and dense *** proposed framework,which employs the U-Net architecture,is straightforward to implement,computationally efficient,and has single-step *** layers are used to retrieve the pooling layers’erased information and learn hierarchically pixelwise spatial *** helps in obtaining feature values,retaining spatial locations,and maximizing data integrity to avoid data *** addition,a modified attention unit is introduced and integrated into the proposed 2UNet model to focus on specific crowd *** proposed model concentrates on balancing the number of model parameters,model size,computational cost,and counting accuracy compared with other works,which may involve acquiring one criterion at the expense of other *** on five challenging datasets for density estimation and crowd counting have shown that the proposed model is very effective and outperforms comparable mainstream ***,it counts very well in both sparse and congested crowd *** 2U-Net model has the lowest MAE in both parts(Part A and Part B)of the ShanghaiTech,UCSD,and Mall benchmarks,with 63.3,7.4,1.5,and 1.6,***,it obtains the lowest MSE in the ShanghaiTech-Part B,UCSD,and Mall benchmarks with 12.0,1.9,and 2.1,respectively.
Artificial Intelligence, including machine learning and deep convolutional neural networks (DCNNs), relies on complex algorithms and neural networks to process and analyze data. DCNNs for visual recognition often requ...
详细信息
Extracting valuable information frombiomedical texts is one of the current research hotspots of concern to a wide range of *** biomedical corpus contains numerous complex long sentences and overlapping relational trip...
详细信息
Extracting valuable information frombiomedical texts is one of the current research hotspots of concern to a wide range of *** biomedical corpus contains numerous complex long sentences and overlapping relational triples,making most generalized domain joint modeling methods difficult to apply effectively in this *** a complex semantic environment in biomedical texts,in this paper,we propose a novel perspective to perform joint entity and relation extraction;existing studies divide the relation triples into several steps or ***,the three elements in the relation triples are interdependent and inseparable,so we regard joint extraction as a tripartite classification *** the same time,fromthe perspective of triple classification,we design amulti-granularity 2D convolution to refine the word pair table and better utilize the dependencies between biomedical word ***,we use a biaffine predictor to assist in predicting the labels of word pairs for relation *** model(MCTPL)Multi-granularity Convolutional Tokens Pairs of Labeling better utilizes the elements of triples and improves the ability to extract overlapping triples compared to previous ***,we evaluated our model on two publicly accessible *** experimental results show that our model’s ability to extract relation triples on the CPI dataset improves the F1 score by 2.34%compared to the current optimal *** the DDI dataset,the F1 value improves the F1 value by 1.68%compared to the current optimal *** model achieved state-of-the-art performance compared to other baseline models in biomedical text entity relation extraction.
暂无评论