检索结果-内蒙古大学图书馆

2023 China Automation Congress, CAC 2023

作者： Zhao, Wei Zhang, Hui Cao, Yihong Liu, Lizhu Su, Yingjian School of Robotics Hunan University Changsha China College of Computer Science and Electronic Engineering Hunan University Changsha China

ISBN: (纸本)9798350303759

Due to the influence of space, time, environmental changes, and types of ground objects on the lighting conditions of high-light falling images, Hyperspectral image (HSI) classification is challenging. Recently, attention-based HSI classification has received increasing attention. It uses both spatial attention and spectral attention to capture the required spatial-spectral features, thereby achieving better classification accuracy. However, it is challenging to connect the two attention mechanisms without compromising the spatial-spectral features. In addition, how to obtain the attention weight matrix which determines the importance of each channel or spatial location is also an important problem in the use of attention mechanisms. In response to the above issues, the Spatial Central Attention-Spectral Attention network (SCASAN) is designed to efficiently model the spatial-spectral information for HSI classification. In this network, we propose the novel spectral weight spatial attention (SWSA) module to solve the connection problem between spatial attention and spectral attention. Furthermore, we utilize a spatial central attention network (spatial CAN) to obtain a spatial attention weight matrix to avoid the impact of non-homogeneous adjacent pixels on the classified pixel. Our proposed approach demonstrates superior performance in hyperspectral image classification compared to state-of-the-art methods, as evidenced by comprehensive experiments conducted on three publicly available datasets. © 2023 IEEE.

关键词： Classification (of information)

来源：评论

学校读者我要写书评

暂无评论

Omnidirectional Autonomous Navigation With Multi-Camera RGBD Odometry 15

Omnidirectional Autonomous Navigation With Multi-Camera RGBD...

引用

15th Annual Undergraduate Research Conference on Applied Computing, URC 2024

作者： Urbanik, Igor Gibiec, Mariusz AGH University of Krakow Faculty of Computer Science Kraków Poland AGH University of Krakow Faculty of Mechanical Engineering and Robotics Kraków Poland

ISBN: (纸本)9798331527341

Omnidirectionality, that is (in the autonomous navigation context), ability to move in any direction at all times without extensive turning provides some nuances that can be leveraged greatly in many scenarios. This paper aims to describe an approach to omnidirectional autonomous navigation and localization based on the Kalman Planetary Rover developed by the AGH Space Systems team at the Faculty of Mechanical Engineering and robotics at AGH University of Krakow. © 2024 IEEE.

关键词： Cameras

来源：评论

学校读者我要写书评

暂无评论

Towards a More Stable and General Subgraph Information Bottleneck 48

Towards a More Stable and General Subgraph Information Bottl...

引用

48th IEEE International Conference on Acoustics, Speech and Signal Processing, ICASSP 2023

作者： Liu, Hongzhi Zheng, Kaizhong Yu, Shujian Chen, Badong Xi'an Jiaotong University Institute of Artifical Intelligience and Robotics China Vrije Universiteit Amsterdam Department of Computer Science Netherlands

ISBN: (纸本)9781728163277

Graph Neural Networks (GNNs) have been widely applied to graph-structured data. However, the lack of interpretability impedes its practical deployment especially in high-risk areas such as medical diagnosis. Recently, the Information Bottleneck (IB) principle has been extended to GNNs to identify a compact subgraph that is most informative to class labels, which significantly improves the interpretability on decision. However, existing Graph Information Bottleneck (GIB) models are either unstable during the training (due to the difficulty of mutual information estimation) or only focus on a special kind of graph (e.g., brain networks) that suffer from poor generalization to general graph datasets with varying graph sizes. In this work, we extend the recently developed Brain Information Bottleneck (BrainIB) to general graphs by introducing matrix-based Rényi's α-order mutual information to stablize the training;and by designing a novel mask strategy to deal with varying graph sizes such that the new method can also be used for social networks, molecules, etc. Extensive experiments on different types of graph datasets demonstrate the superior stability and generality of our model. © 2023 IEEE.

关键词： Generalization Graph Information Bottleneck Stability

来源：评论

学校读者我要写书评

暂无评论

Bridging the gap between Learning-to-plan, Motion Primitives and Safe Reinforcement Learning 8

Bridging the gap between Learning-to-plan, Motion Primitives...

引用

8th Conference on Robot Learning, CoRL 2024

作者： Kicki, Piotr Tateo, Davide Liu, Puze Guenster, Jonas Peters, Jan Walas, Krzysztof IDEAS NCBR Warsaw Poland Institute of Robotics and Machine Intelligence Poznan University of Technology Poland Department of Computer Science Technische Universitat Darmstadt Germany

Trajectory planning under kinodynamic constraints is fundamental for advanced robotics applications that require dexterous, reactive, and rapid skills in complex environments. These constraints, which may represent task, safety, or actuator limitations, are essential for ensuring the proper functioning of robotic platforms and preventing unexpected behaviors. Recent advances in kinodynamic planning demonstrate that learning-to-plan techniques can generate complex and reactive motions under intricate constraints. However, these techniques necessitate the analytical modeling of both the robot and the entire task, a limiting assumption when systems are extremely complex or when constructing accurate task models is prohibitive. This paper addresses this limitation by combining learning-to-plan methods with reinforcement learning, resulting in a novel integration of black-box learning of motion primitives and optimization. We evaluate our approach against state-of-the-art safe reinforcement learning methods, showing that our technique, particularly when exploiting task structure, outperforms baseline methods in challenging scenarios such as planning to hit in robot air hockey. This work demonstrates the potential of our integrated approach to enhance the performance and safety of robots operating under complex kinodynamic constraints. © 2024 Proceedings of Machine Learning Research.

关键词： Reinforcement learning

来源：评论

学校读者我要写书评

暂无评论

Deep Reinforcement Learning for Solving the Vehicle Routing Problem in Practical Logistics 13

Deep Reinforcement Learning for Solving the Vehicle Routing ...

引用

13th IEEE Congress on Evolutionary Computation, CEC 2024

作者： Cai, Junchuang Zhang, Xinzhi Lin, Qiuzhen Dong, Lisha Chen, Weineng Ming, Zhong Shenzhen University College of Computer Science and Software Engineering Shenzhen518060 China Shenzhen Institute of Information Technology College of Sino-German Robotics Shenzhen China SSouth China University of Technology College of Computer Science and Engineering Guangzhou China

ISBN: (纸本)9798350308365

The vehicle routing problem with simultaneous pickup-delivery and time windows (VRPPDTW) is applicable to a wide range of practical scenarios within the domains of transportation and logistics. When addressing this complex problem, conventional heuristic algorithms have exhibited limitations in effectiveness, primarily due to the need for expert knowledge in heuristic design, which often prioritizes immediate improvements and leads to convergence towards local optima. To overcome this challenge, we propose a deep Q-network-based iterated local search algorithm (DQN-ILS). This algorithm employs a deep Q-learning network to automatically learn a policy for selecting heuristics to improve the current solution. Simultaneously, an adaptive perturbation strategy is designed to avoid becoming trapped in local optima. In this way, our algorithm can better balance the trade-off between exploration and exploitation when solving VRPPDTW. Experimental studies conducted on a real test suite of large-scale VRPPDTW instances validate the superiority of the proposed DQN-ILS over several recently proposed methods. © 2024 IEEE.

关键词： Deep reinforcement learning

来源：评论

学校读者我要写书评

暂无评论

A Residual Attention-based EfficientNet Homography Estimation Model for Sports Field Registration 49

A Residual Attention-based EfficientNet Homography Estimatio...

引用

49th Annual Conference of the IEEE Industrial Electronics Society, IECON 2023

作者： Oo, Yin May Jamsrandorj, Ankhzaya Chao, Vanyi Mun, Kyung-Ryoul Kim, Jinwook Ai Robotics University of Science and Technology Daejeon34113 Korea Republic of Human Computer Interface and Robotics University of Science and Technology Daejeon34113 Korea Republic of Center for Artificial Intelligence Korea Institute of Science and Technology Seoul02792 Korea Republic of

ISBN: (纸本)9798350331820

Accurate sports field registration plays a crucial role in computer vision for effective team sports analysis, as it provides the ability to track and recognize the positions, movements, and interactions of players from image space to real-world field space. In this paper, we propose a two-stage deep learning framework for sports field registration. In the first-stage network, we utilize a Residual EfficientNet-Attention U-Net architecture to estimate the initial homography matrix using predefined keypoints to register sports fields. Subsequently, our second-stage network refines the initial homography for improved accuracy. We evaluate our proposed method on the public sports datasets to show its outperform results against the state-of-the-art methods in some evaluation metrics. © 2023 IEEE.

关键词： Sports

来源：评论

学校读者我要写书评

暂无评论

3D Ball Trajectory Reconstruction of a Ballistic Shot from a Monocular Basketball Video 49

3D Ball Trajectory Reconstruction of a Ballistic Shot from a...

引用

49th Annual Conference of the IEEE Industrial Electronics Society, IECON 2023

作者： Chao, Vanyi Jamsrandorj, Ankhzaya Oo, Yin May Mun, Kyung-Ryoul Kim, Jinwook University of Science and Technology Ai Robotics Daejeon34113 Korea Republic of University of Science and Technology Human Computer Interface and Robotics Daejeon34113 Korea Republic of Center for Artificial Intelligence Korea Institute of Science and Technology Seoul02792 Korea Republic of

ISBN: (纸本)9798350331820

3D ball location in sports can be advantageous and has many applications such as automatic scoring system, ball trajectory highlight and offside detection system for football game. Currently, this task can be solved through the use of costly multi-view approaches or the legacy physics-based methods that require complicated setups and human intervention. In this paper, we propose a method to reconstruct the 3D ball trajectory of the basketball projectile shot by using only a single view. The method relies on accurate camera calibration and 2D ball detection of sequence images. Extra information in the 3D world also plays an important role in improving the 3D reconstruction. The experimental result illustrates the accurately reconstructed 3D trajectory when enough information is given. © 2023 IEEE.

关键词： Ballistics

来源：评论

学校读者我要写书评

暂无评论

V-TX: A VR-Based Deployment and Evaluation System for Wearable Telexistence Robotic Systems 23

V-TX: A VR-Based Deployment and Evaluation System for Wearab...

引用

23rd International Conference on Cyberworlds, CW 2024

作者： Iskandar, Abdullah Aburajouh, Hala Khazer Shebli Al-Shaer, Eman Hussain Al-Yafei, Fatima Thabit Al-Arqaban, Bushra Ahmad Al-Yousef, Moneera Yang, Pinchu Halabi, Osama Al-Sada, Mohammed Nakajima, Tatsuo Waseda University Computer Science and Engineering Japan Telkom University School of Computing Indonesia Qatar University Computer Science and Engineering Qatar Waseda University Future Robotics Organization Japan Qatar University KINDI Computing Research Center Qatar

ISBN: (纸本)9798331527174

Wearable telexistence robotic systems have risen in prominence after the COVID-19 pandemic, being vital for various applications. However, their design, fabrication, and evaluation require substantial resources and are unsupported by robot simulation systems. Therefore, we present V-TX, which is a VR-based system for deploying and evaluating wearable telexistence systems. V-TX extends key telexistence systems' components with versatility for customizing different attributes based on the intended application. To demonstrate the potential of V-TX, we replicated a real telexistence robot in V-TX, called Piton, and showed a novel potential use for exploring and evaluating different robot designs, interaction methods, and deployments. We conducted a comparative evaluation between the real Piton robot and its V-TX implementation. The results show that V-TX gave results that match those of the real counterpart across the main factors of telexistence, indicating its viability as an alternative to traditional evaluation approaches using real robots. Lastly, we discuss the limitations and present future research directions. © 2024 IEEE.

关键词： Machine design

来源：评论

学校读者我要写书评

暂无评论

How Language of Interaction Affects the User Perception of a Robot 15th

How Language of Interaction Affects the User Perception of...

引用

15th International Conference on Social robotics, ICSR 2023

作者： Sienkiewicz, Barbara Sejnova, Gabriela Gajewski, Paul Vavrecka, Michal Indurkhya, Bipin Cognitive Science Department Jagiellonian University Krakow Poland Czech Institute of Informatics Robotics and Cybernetics Czech Technical University Prague Czech Republic Institute of Computer Science AGH University of Science and Technology Krakow Poland

ISBN: (纸本)9789819987146

Spoken language is the most natural way for a human to communicate with a robot. It may seem intuitive that a robot should communicate with users in their native language. However, it is not clear if a user’s perception of a robot is affected by the language of interaction. We investigated this question by conducting a study with twenty-three native Czech participants who were also fluent in English. The participants were tasked with instructing the Pepper robot on where to place objects on a shelf. The robot was controlled remotely using the Wizard-of-Oz technique. We collected data through questionnaires, video recordings, and a post-experiment feedback session. The results of our experiment show that people perceive an English-speaking robot as more intelligent than a Czech-speaking robot (z = 18.00, p-value = 0.02). This finding highlights the influence of language on human-robot interaction. Furthermore, we discuss the feedback obtained from the participants via the post-experiment sessions and its implications for HRI design. © 2024, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.

关键词： Feedback

来源：评论

学校读者我要写书评

暂无评论

Data Augmentation via Face Morphing for Recognizing Intensities of Facial Emotions

引用

IEEE Transactions on Affective Computing 2023年第2期14卷 1228-1235页

作者： Huang, Tsung-Ren Hsu, Shin-Min Fu, Li-Chen National Taiwan University Department of Psychology Center for Artificial Intelligence & Advanced Robotics Taipei106319 Taiwan National Taiwan University Center for Artificial Intelligence & Advanced Robotics Department of Electrical Engineering Department of Computer Science and Information Engineering Taipei106319 Taiwan

Being able to recognize emotional intensity is a desirable feature for a facial emotional recognition (FER) system. However, the development of such a feature is hindered by the paucity of intensity-labeled data for model training. To ameliorate the situation, the present study proposes using face morphing as a novel way of data augmentation to synthesize faces that express different degrees of a designated emotion. Such an approach has been successfully validated on humans and machines. Specifically, humans indeed perceived different levels of intensified emotions in these parametrically synthesized faces, and FER systems based on neural networks indeed showed improved sensitivities to intensities of different emotions when additionally trained on the synthesized faces. Overall, the proposed data augmentation method is not only simple and effective but also useful for building FER systems that recognize facial expressions of mixed emotions. © 2010-2012 IEEE.

关键词： Face recognition

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：