检索结果-内蒙古大学图书馆

Continuous Geodesic Self-Attention Models with Gated Fusion for Trajectory Prediction

学校读者我要写书评

暂无评论

Continuous Geodesic Self-Attention Models with Gated Fusion ...

International Joint Conference on Neural Networks (IJCNN)

作者： Kexin Ke Zhengyu Li Huining Chen Hao Wang Xian Wei Jian Yang Xuan Tang MoE Engineering Research MOE Center of SW/HW Co-Design Technology and Application Shanghai Key Laboratory of Trustworthy Computing East China Normal University China School of Advanced Technology Xi’an Jiaotong-Liverpool University China School of Geospatial Information Information Engineering University China School of Communications and Electronic Engineering East China Normal University China

ISBN: (数字)9798350359312

ISBN: (纸本)9798350359329

Driven by the rapid development of intelligent driving vehicles, predicting the trajectories of pedestrians on the road is crucial for decision-making during driving and even road safety. In this paper, we propose a novel method for trajectory prediction, namely, Continuous Geodesic Self-Attention Models with Gated Fusion (CGSAG). We use geodesic attention to measure the similarity between trajectory points, and utilize a gating mechanism to fuse the geodesic features extracted by multi-layer graph convolution. We then use Neural Ordinary Differential Equations (Neural ODE) to model the continuous-time dynamics of the trajectory. We show that CGSAG improves state-of-the-art performances on several human trajectory prediction datasets, including ETH/UCY, SDD, and Ind. At the same time, we conduct ablation studies to prove the effectiveness and efficiency of our proposed method.

关键词： Pedestrians Logic gates Predictive models Feature extraction Mathematical models Time measurement Road safety

Stfi: Structure-Aware and Time-Frequency Domain Interaction Network for Deformable Medical Image Registration

学校读者我要写书评

暂无评论

SSRN 2025年

作者： Hu, Rui Wang, Nian Zhang, Yuxin Li, Teng Key Laboratory of Intelligent Computing and Signal Processing Ministry of Education School of Electrical Engineering and Automation Anhui University Hefei230601 China Advanced Imaging Research Center University of Texas Southwestern Medical Center 5323 Harry Hines Blvd Dallas75390 United States School of Cyber Science and Engineering Southeast University No.2 Southeast University Road Nanjing211189 China

Deformable image registration is crucial in medical image analysis. Recently, unsupervised methods using intensity matching have shown promising results without labeled data. However, these methods still face challenges: (1) They often encounter local optima when handling anatomical structures with significant variations in intensity distribution;(2) Time domain convolution in decoders leads to loss of high-frequency fine features, limiting the effectiveness of generating fine deformation vector fields. To address these issues, we propose a novel structure-aware and time-frequency domain interaction (STFI) network. STFI integrates a structure-aware frequency domain loss and a time-frequency domain interaction module. The former leverages principal structural information from the weighted phase spectrum to guide the registration process, thereby reducing the risk of local optima. The latter processes and amalgamates information from both the time and frequency domains, effectively preserving high-frequency details. The efficacy and adaptability of our approach have been demonstrated on two types of data, offering a new perspective on exploiting time-frequency domain information for medical image registration. © 2025, The Authors. All rights reserved.

关键词： Image registration

Finite-time asynchronous dissipative filtering of conic-type nonlinear Markov jump systems

学校读者我要写书评

暂无评论

Science China(Information Sciences) 2021年第5期64卷 151-162页

作者： Xiang ZHANG Shuping HE Vladimir STOJANOVIC Xiaoli LUAN Fei LIU Key Laboratory of Intelligent Computing and Signal Processing (Ministry of Education) School of Electrical Engineering and Automation Anhui University Department of Automatic Control Robotics and Fluid Technique Faculty of Mechanical and Civil EngineeringUniversity of Kragujevac Key Laboratory of Advanced Process Control for Light Industry (Ministry of Education) Institute of AutomationJiangnan University

In the present study, the finite-time asynchronous dissipative filter design problem for the Markov jump systems with conic-type nonlinearity is studied. The hidden Markov model can describe the asynchronism embodied in the system modes and the filter modes reasonably. Moreover, a suitable LyapunovKrasovskii function is utilized and linear matrix inequalities are applied to obtain adequate conditions. These techniques guarantee the finite-time boundedness and strict dissipativity of the filtering error dynamic system. Furthermore, the design problems of the passive filter and the H∞ filter are studied by adjusting the three parameters U, G and V. Finally, the filter gains and the optimal index α*are obtained and the correctness and feasibility of the designed approach are verified by a simulation example.

关键词： dissipative filtering hidden Markov model finite-time boundedness Markov jump systems conic-type nonlinearity

Thermal Error Prediction for Electrical Spindles Based on GA-GRNN

学校读者我要写书评

暂无评论

SSRN 2024年

作者： Li, Yuan Liu, Xiao Dai, Ye Zhan, ShiQiang Wang, JianHui Wang, ZiJian Wan, ShiMing School of Mechanical Engineering and Automation University of Science and Technology Liaoning Anshan114051 China Key Laboratory of Advanced Manufacturing Intelligent Technology Ministry of Education Harbin University of Science and Technology Harbin150080 China Institute of Digital Design and Automatic Machinery Product Development Harbin University of Science and Technology Harbin150080 China Beijing Institute of Technology China

A significant way to increase the spindle's machining precision is by real-time compensation of thermal error in machine coordinates. The accuracy and versatility of the thermal error prediction model, which serves as the foundation for compensation, are very crucial. This paper aims to increase the accuracy of the thermal error prediction model by first calculating and analyzing the thermal characteristics of the electrical spindle, and then using the finite element method to obtain the distribution law of the temperature field of the electrical spindle;the number of temperature measurement points is optimized to three in order to reduce redundancy in the thermal error modeling data;finally, the hybrid algorithm thermal error prediction model is created with the temperature data of the electrical spindle as the input variable and the thermal error data as the output variable. The hybrid algorithm prediction model offers a new reference technique for the study of thermal error compensation of electrical spindles and has the advantages of high accuracy, high speed, and good generalization ability when compared to the radial basis function neural network algorithm. © 2024, The Authors. All rights reserved.

关键词： Temperature measurement

Explore Contextual Information for 3D Scene Graph Generation

学校读者我要写书评

暂无评论

arXiv

arXiv 2022年

作者： Liu, Yuanyuan Long, Chengjiang Zhang, Zhaoxuan Liu, Bokai Zhang, Qiang Yin, Baocai Yang, Xin The Department of Electronic Information and Electrical Engineering Dalian University of Technology Dalian116024 China Meta Reality Labs BurlingameCA94010 United States The Key Lab of Advanced Design and Intelligent Computing Dalian University Ministry of Education Dalian116622 China The Beijing Key Laboratory of Multimedia and Intelligent Software Technology Beijing Institute of Artificial Intelligence Faculty of Information Technology Beijing University of Technology Beijing100124 China

3D scene graph generation (SGG) has been of high interest in computer vision. Although the accuracy of 3D SGG on coarse classification and single relation label has been gradually improved, the performance of existing works is still far from being perfect for fine-grained and multi-label situations. In this paper, we propose a framework fully exploring contextual information for the 3D SGG task, which attempts to satisfy the requirements of fine-grained entity class, multiple relation labels, and high accuracy simultaneously. Our proposed approach is composed of a Graph Feature Extraction module and a Graph Contextual Reasoning module, achieving appropriate information-redundancy feature extraction, structured organization, and hierarchical inferring. Our approach achieves superior or competitive performance over previous methods on the 3DSSG dataset, especially on the relationship prediction sub-task. Copyright © 2022, The Authors. All rights reserved.

关键词： Feature extraction

Effect of the Graphene Oxide Incorporation on the Morphology, Corrosion Resistance and Fatigue Performance of Micro-Arc Oxidation Coating on Magnesium Alloy

学校读者我要写书评

暂无评论

SSRN 2024年

作者： Tan, Yansong Liu, Xin Sun, Xiaohao Gao, Lilan Liu, Debao Zhang, Chunqiu Tianjin Key Laboratory for Advanced Mechatronic System Design and Intelligent Cont rol School of Mechanical Engineering Tianjin University of Technology Tianjin300384 China National Demonstration Center for Experimental Mechanical and Electrical Engineerin g Education Tianjin University of Technology Tianjin300384 China School of Materials Science and Engineering Tianjin University of Technology Tianjin300384 China

Magnesium (Mg) alloy has great potential in biomedical applications, but the rapid corrosion rate which mismatches with bone reconstruction rate limits its clinical use. In this study, the graphene oxides (GO) were added into alkaline electrolyte solution during micro-arc oxidation (MAO) process, aiming to form a corrosion resistant surface coating on Mg alloy. By changing the concentration of GO, the coatings' morphology, composition, corrosion resistance, and fatigue property were compared to reveal the strengthening mechanism and find out the optimal graphene concentration. The results show that the MAO coating is composed of MgO, Mg2SiO4, Mg3(PO4)2 and GO. The added GO exhibits excellent adsorption property in the MAO process, reducing the intermolecular gap, thereby improving densification and obturation of the MAO coating. The surface morphology of MAO coating initially improves with increasing GO content, but then deteriorates, reaching its best quality at 0.4 g/L. Compared to MAO samples without GO, the samples with 0.4g/L GO not only have lower corrosion current density of 5.226 μA/cm2, but also show higher fatigue lives both before and after pre-corrosion. This study offers a promising approach to enhance the corrosion resistance and fatigue performance of Mg alloy by applying a GO-strengthened MAO coating onto its surface. © 2024, The Authors. All rights reserved.

关键词： Magnesium alloys

Accurate MTPA Control Based on Ferrari Method for IPMSM with Partial Derivative Term Calculation

学校读者我要写书评

暂无评论

Accurate MTPA Control Based on Ferrari Method for IPMSM with...

International Conference on Electrical Machines and Systems (ICEMS)

作者： Yiyang Li Xin Gu Zhiqiang Wang Liyan Guo Wei Chen Xinmin Li Chen Li School of Electronics and Information Engineering Tiangong University Tianjin China Tianjin Key Laboratory of Intelligent Control of Electrical Equipment Tianjin China School of Electrical Engineering Tiangong University Tianjin China National Local Joint Engineering Research Center of Electrical Machine System Design and Manufacturing China Zhejiang University Advanced Electrical Equipment Innovation Center Hangzhou China

ISBN: (数字)9784886864406

ISBN: (纸本)9798350379105

Even if the exact motor parameters are known, the MTPA strategy cannot obtain precise maximum torque per ampere (MTPA) points under large current amplitude because the partial derivative of inductance to current is ignored in the modeling process. In view of this, the MTPA mathematical model considering the partial derivative of inductance to current is established in this paper, and the MTPA point tracking error caused by ignoring the partial derivative in the modeling process based on Ferrari method is analyzed. Then, an offline-online combined method for calculating the partial derivative term is proposed to compensate the MTPA tracking error. On this basis, an MTPA control strategy based on Ferrari method considering the partial derivative of inductance to current is proposed. Experiment to verify the feasibility and effectiveness of the proposed method.

关键词： Inductance Analytical models Torque Accuracy Permanent magnet motors Mathematical models Trajectory

Effect of the Graphene Oxide Incorporation on the Morphology, Corrosion Resistance And Fatigue Performance Of Micro-Arc Oxidation Coating on Mg Alloy

学校读者我要写书评

暂无评论

SSRN 2024年

Mg alloy has great potential in biomedical applications, but the rapid corrosion rate which mismatches with bone reconstruction rate limits its clinical use. A protective oxide coating layer can be generated on the surface of Mg alloy by micro-arc oxidation (MAO) technique, aiming to improve the corrosion performance and long-term structural integrity of Mg alloy. In this study, the MAO coating layer of Mg alloy was improved by adding appropriate content of graphene oxide (GO) into electrolyte solution. GO has excellent adsorption property and plays a role in the MAO process, reducing the intermolecular gap, thereby improving the densification and obturation of the coating surface. With the increase of GO content, the surface morphology of the MAO coating initially has improved, then later deteriorated, and the best surface morphology has achieved at a GO content of 0.4 g/L. When the GO content increases from 0 g/L to 0.4 g/L, the corrosion current density reduces to 5.226 μA/cm2, with a reduction two orders in magnitude;in addition, at a stress amplitude of 140 MPa and a frequency of 4 Hz, the fatigue life increases from 9261 cycles to 13887 cycles without corrosion, and from 2042 cycles to 2759 cycles after pre-corrosion. This study provided a promising approach to ameliorate the morphology, corrosion resistance and fatigue performance of MAO coating on Mg alloy by elaborately configuring the electrolyte solution and customizing the process parameters. © 2024, The Authors. All rights reserved.

关键词： Corrosion resistance

Low-Light Enhancement Effect on Classification and Detection: An Empirical Study

学校读者我要写书评

暂无评论

arXiv

arXiv 2024年

作者： Wu, Xu Lai, Zhihui Jie, Zhou Gao, Can Hou, Xianxu Zhang, Ya-Nan Shen, Linlin The Computer Vision Institute College of Computer Science and Software Engineering Shenzhen University Shenzhen518060 China Shenzhen Institute of Artificial Intelligence and Robotics for Society Shenzhen518060 China Guangdong Key Laboratory of Intelligent Information Processing Shenzhen University Shenzhen518060 China School of AI and Advanced Computing Xi’an Jiaotong-Liverpool University China The National Engineering Laboratory for Big Data System Computing Technology Shenzhen University SZU Branch Shenzhen Institute of Artificial Intelligence and Robotics for Society Guangdong Shenzhen518060 China

Low-light images are commonly encountered in real-world scenarios, and numerous low-light image enhancement (LLIE) methods have been proposed to improve the visibility of these images. The primary goal of LLIE is to generate clearer images that are more visually pleasing to humans. However, the impact of LLIE methods in high-level vision tasks, such as image classification and object detection, which rely on high-quality image datasets, is not well explored. To explore the impact, we comprehensively evaluate LLIE methods on these high-level vision tasks by utilizing an empirical investigation comprising image classification and object detection experiments. The evaluation reveals a dichotomy: While Low-Light Image Enhancement (LLIE) methods enhance human visual interpretation, their effect on computer vision tasks is inconsistent and can sometimes be harmful. Our findings suggest a disconnect between image enhancement for human visual perception and for machine analysis, indicating a need for LLIE methods tailored to support high-level vision tasks effectively. This insight is crucial for the development of LLIE techniques that align with the needs of both human and machine vision. Copyright © 2024, The Authors. All rights reserved.

关键词： Machine vision