Point clouds are naturally sparse, while image pixels are dense. The inconsistency limits feature fusion from both modalities for point-wise scene flow estimation. Previous methods rarely predict scene flow from the e...
Point clouds are naturally sparse, while image pixels are dense. The inconsistency limits feature fusion from both modalities for point-wise scene flow estimation. Previous methods rarely predict scene flow from the entire point clouds of the scene with one-time inference due to the memory inefficiency and heavy overhead from distance calculation and sorting involved in commonly used farthest point sampling, KNN, and ball query algorithms for local feature aggregation. To mitigate these issues in scene flow learning, we regularize raw points to a dense format by storing 3D coordinates in 2D grids. Unlike the sampling operation commonly used in existing works, the dense 2D representation 1) preserves most points in the given scene, 2) brings in a significant boost of efficiency, and 3) eliminates the density gap between points and pixels, allowing us to perform effective feature fusion. We also present a novel warping projection technique to alleviate the information loss problem resulting from the fact that multiple points could be mapped into one grid during projection when computing cost volume. Sufficient experiments demonstrate the efficiency and effectiveness of our method, outperforming the prior-arts on the FlyingThings3D and KITTI dataset. Our source codes will be released on https://***/IRMVLab/DELFlow.
Multiple types of sensors are utilized for remote monitoring of single industrial process to compensate for the limited accuracy of a single type of sensor. Network slicing technology is considered promising for colle...
详细信息
A fuzzy creative works generation algorithm based on graph neural network is proposed. Firstly, the multi-label fuzzy creative data set is constructed. Secondly, fuzzy logical correlations between creative objects are...
详细信息
Aiming at accurate identification of emerging new faults in PV systems, a novel PV fault diagnosis framework based on cloud-edge collaboration is designed. Model updates are deployed on the cloud servers, while real-t...
详细信息
In this research,a novel dynamic and heterogeneous identity based cooperative co-evolutionary algorithm(DHICCA)is proposed for addressing the distributed lot-streaming flowshop scheduling problem(DLSFSP)with the objec...
详细信息
In this research,a novel dynamic and heterogeneous identity based cooperative co-evolutionary algorithm(DHICCA)is proposed for addressing the distributed lot-streaming flowshop scheduling problem(DLSFSP)with the objective to minimize the makespan.A two-layer-vector representation is devised to bridge the solution space of DLSFSP and the search space of *** the evolution of DHICCA,population individuals are endowed with heterogeneous identities according to their quality,including superior individuals,ordinary individuals,and inferior individuals,which serve local exploitation,global exploration,and diversified restart,*** individuals with different identities require different evolutionary mechanisms to fully unleash their respective potentials,identity-specific evolutionary operators are devised to evolve them in a cooperative co-evolutionary *** is important to use limited population resources to solve complex optimization ***,exploitation is carried out on superior individuals by devising three exploitative operators with different intensities based on techniques of variable neighborhood,destruction-construction,and gene *** is executed on ordinary individuals by a newly constructed discrete Jaya algorithm and a probability crossover *** addition,restart is performed on inferior individuals to introduce new evolutionary individuals to the *** the cooperative co-evolution,all individuals with different identities are merged as a population again,and their identities are dynamically adjusted by new *** influence of parameters on the algorithm is investigated based on design-of-experiment and comprehensive computational experiments are used to evaluate the performance of all *** results validate the effectiveness of special designs and show that DHICCA performs more efficient than the existing state-of-the-art algorithms in solving the DLSFSP.
Receiver plays a vital role in various electronic countermeasure systems and has always been a hot topic to research. Based on the traditional receiver digital channelization structure, this paper further researches a...
详细信息
In this work, we experimentally investigated the gain characteristics of a ring-core erbium-doped fiber (RC-EDF) that supports five orbital angular momentum (OAM) mode groups. Through the design and successful fabrica...
详细信息
Electrostatic dielectric capacitors with ultrahigh power densities are sought after for advanced electronic and electrical systems owing to their ultrafast charge-discharge capability. However, low energy density resu...
Electrostatic dielectric capacitors with ultrahigh power densities are sought after for advanced electronic and electrical systems owing to their ultrafast charge-discharge capability. However, low energy density resulting from low breakdown strength and suppressed polarization still remains a daunting challenge for practical applications. We propose a microstructural strategy with dendritic nanopolar (DNP) regions self-assembled into an insulator, which simultaneously enhances breakdown strength and high-field polarizability and minimizes energy loss and thus markedly improves energy storage performance and stability. For illustration, in this study, we achieved a high energy density of 215.8 joules per cubic centimeter with an efficiency of 80.7% at a high electric field of 7.4 megavolts per centimeter in a DNP structure-designed PbZr0.53Ti0.47O3-MgO film. The proposed strategy is generally applicable for development of high-performance dielectric microcapacitors.
We developed a mixed-strategy gaming approach for distributed and real-time optical subscarrier management in point-to-multipoint networks, achieving traffic loss rates close to those by ILP-based centralized optimiza...
作者:
He, ChaoYang, NanLeung, Kwok WaLiaoning Technical University
Liaoning Key Laboratory of Radio Frequency and Big Data for Intelligent Applications Huludao125105 China Sun Yat-sen University
School of Electronics and Information Technology Guangdong Provincial Key Laboratory of Optoelectronic Information Processing Chips and Systems Guangzhou China City University of Hong Kong
State Key Laboratory of Terahertz and Millimeter Waves Department of Electrical Engineering Kowloon Hong Kong CityU Shenzhen Research Institute
Information and Communication Technology Centre Shenzhen China
A four-port multi-input multi-output (MIMO) dielectric resonator (DR) antenna (DRA) is presented in this paper. Combing four rectangular DRs together can form a single DR element. With shape modification and conformal...
详细信息
暂无评论