检索结果-内蒙古大学图书馆

Hourglass-GCN for 3D Human Pose Estimation Using Skeleton Structure and View Correlation

Computers, Materials & Continua 2025年第1期82卷 173-191页

作者： Ange Chen Chengdong Wu Chuanjiang Leng Faculty of Robot Science and Engineering Northeastern UniversityShenyang110169China

Previous multi-view 3D human pose estimation methods neither correlate different human joints in each view nor model learnable correlations between the same joints in different views explicitly,meaning that skeleton structure information is not utilized and multi-view pose information is not completely ***,existing graph convolutional operations do not consider the specificity of different joints and different views of pose information when processing skeleton graphs,making the correlation weights between nodes in the graph and their neighborhood nodes *** Graph Convolutional Networks(GCNs)cannot extract global and deeplevel skeleton structure information and view correlations *** solve these problems,pre-estimated multiview 2D poses are designed as a multi-view skeleton graph to fuse skeleton priors and view correlations explicitly to process occlusion problem,with the skeleton-edge and symmetry-edge representing the structure correlations between adjacent joints in each viewof skeleton graph and the view-edge representing the view correlations between the same joints in different *** make graph convolution operation mine elaborate and sufficient skeleton structure information and view correlations,different correlation weights are assigned to different categories of neighborhood nodes and further assigned to each node in the *** on the graph convolution operation proposed above,a Residual Graph Convolution(RGC)module is designed as the basic module to be combined with the simplified Hourglass architecture to construct the Hourglass-GCN as our 3D pose estimation ***-GCNwith a symmetrical and concise architecture processes three scales ofmulti-viewskeleton graphs to extract local-to-global scale and shallow-to-deep level skeleton features *** results on common large 3D pose dataset Human3.6M and MPI-INF-3DHP show that Hourglass-GCN outperforms some excellent methods in 3D pose estimati

关键词： 3D human pose estimation multi-view skeleton graph elaborate graph convolution operation Hourglass-GCN

来源：评论

学校读者我要写书评

暂无评论

Finite-Time Event-Triggered Containment Maneuvering of Marine Surface Vehicles With Tracking Error Constraints: Theory and Experiment

IEEE Transactions on Intelligent Vehicles

引用

IEEE Transactions on Intelligent Vehicles 2024年 1-11页

作者： Feng, Yuxi Wang, Hao Fu, Jun Faculty of Robot Science and Engineering Northeastern University Shenyang China

In this article, a novel class of containment maneuvering controllers is developed for a fleet of marine surface vehicles (MSVs) capable of tracking a parameterized path within finite time. Besides, the along-track errors and cross-track errors of parameterized line-of-sight (LOS) are constrained. For kinematic level, to address tracking error constraints and finite time convergence, the coordinated guidance laws are derived by developing LOS based tan-type barrier Lyapunov functions (BLFs), nonlinear tracking differentiator (NLTD) and finite-time theory. Specifically, a novel finite-time extended state observer (FTESO) is developed to estimate the uncertain kinematics. For coordination level, the finite-time switching threshold event-triggered containment control strategy for MSV is derived to reduce the updating frequency of the path variables containment controller. Theoretical analysis illustrates that all signals are ultimately uniformly bounded (UUB). Both simulation and experimental results are provided to verify the effective performance of the proposed containment maneuvering strategy. IEEE

关键词： Kinematics

来源：评论

学校读者我要写书评

暂无评论

HyGFNet: Hybrid Geometry-Flow Learning Network for 3D Single Object Tracking

IEEE Transactions on Intelligent Vehicles

引用

IEEE Transactions on Intelligent Vehicles 2024年第10期9卷 1-12页

作者： Cui, Yubo Fang, Zheng Li, Zhiheng Li, Shuo Lin, Yu Faculty of Robot Science and Engineering Northeastern University Shenyang China

3D single object tracking (SOT) which attempts to accurately locate the target object in the current frame, has made significant advancements over past years. However, most previous works built upon the Siamese architecture usually focus on the learning and matching of geometry information, while neglecting the motion information of the target. Consequently, those methods face challenges when distinguishing the target object from similar distractors. To overcome this limitation, we enhance the architecture by incorporating flow estimation, presenting a novel multi-frame hybrid geometry-flow learning network for 3D SOT. The proposed framework exploits both geometry and flow information from historical frames and further integrates the learned information into the current frame to improve the target object localization. Specifically, we propose a geometry matching branch and a flow estimation branch. The geometry matching branch first captures the geometry feature of each frame and then aligns these features through multi-frame spatial-temporal matching, leading to a geometry-aware feature for the target object. Meanwhile, in the flow estimation branch, a multi-frame flow-aware enhancement module is proposed to explicitly capture flow information across frames, leading to a flow-aware feature for the target object. Finally, the geometry-aware and flow-aware features are fused with the original feature to predict the position of the target object. Extensive experiments conducted on the KITTI and nuScenes datasets validate the effectiveness and show the competitive performance of our method. The code will be open soon. IEEE

关键词： Geometry

来源：评论

学校读者我要写书评

暂无评论

Brain Cognition Mechanism-Inspired Hierarchical Navigation Method for Mobile robots

引用

Journal of Bionic engineering 2024年第2期21卷 852-865页

作者： Qiang Zou Chengdong Wu Ming Cong Dong Liu Faculty of Robot Science and Engineering Northeastern UniversityShenyang110819China School of Mechanical Engineering Dalian University of TechnologyDalian116024China

Autonomous navigation is a fundamental problem in *** methods generally build point cloud map or dense feature map in perceptual space;due to lack of cognition and memory formation mechanism,traditional methods exist poor robustness and low cognitive *** a new navigation technology that draws inspiration from mammal’s navigation,bionic navigation method can map perceptual information into cognitive space,and have strong autonomy and environment *** improve the robot’s autonomous navigation ability,this paper proposes a cognitive map-based hierarchical navigation ***,the mammals’navigation-related grid cells and head direction cells are modeled to provide the robots with location *** then a global path planning strategy based on cognitive map is proposed,which can anticipate one preferred global path to the target with high efficiency and short ***,a hierarchical motion controlling method is proposed,with which the target navigation can be divided into several sub-target navigation,and the mobile robot can reach to these sub-targets with high confidence ***,some experiments are implemented,the results show that the proposed path planning method can avoid passing through obstacles and obtain one preferred global path to the target with high efficiency,and the time cost does not increase extremely with the increase of experience nodes *** motion controlling results show that the mobile robot can arrive at the target successfully only depending on its self-motion information,which is an effective attempt and reflects strong bionic properties.

关键词： Bionic navigation Spatial localization cells Global path planning Hierarchical motion controlling

来源：评论

学校读者我要写书评

暂无评论

Skeleton-Based Action Recognition with Adaptive Connections and Motion Modeling 6

Skeleton-Based Action Recognition with Adaptive Connections ...

引用

6th International Conference on Industrial Artificial Intelligence, IAI 2024

作者： Geng, Qizhao Zheng, Chao Li, Ziao Wang, Junyi Northeastern University Faculty of Robot Science and Engineering Shenyang China Northeastern University Faculty of Robot Science and Engineering and Foshan Graduate School of Innovation China

ISBN: (纸本)9798350356618

In the field of skeleton-based action recognition, ST-GCN has gained popularity because of its strong capability to model graph data and its effective use of prior knowledge regarding human joint connections. However, it faces certain challenges and limitations, such as difficulties in transferring information between non-adjacent joints and dealing with subtle variations in action speed and acceleration. To address these issues, we present a Graph Convolution Network based Adaptive connections and Motion modeling (AM-GCN). Firstly, we remove the constraints of skeleton graphs by using learnable adjacency matrices to represent dynamic connections between joints, resulting in information exchange between non-adjacent joints. Additionally, we propose a joint motion modeling technique that combines position, velocity, and acceleration information to provide richer prior knowledge about human movements, further enhancing the model's performance. Extensive experiments conducted on the NTU-RGB+D dataset show that our method performs well in skeleton-based human behavior analysis tasks, demonstrating its potential and value in practical applications. © 2024 IEEE.

关键词： Knowledge graph

来源：评论

学校读者我要写书评

暂无评论

Leveraging Frequency and Spatial Domain Information for Underwater Image Restoration

Leveraging Frequency and Spatial Domain Information for Unde...

引用

2024 International Conference on Intelligent Systems and robotics, CISR 2024

作者： Zhang, Haopeng Xu, Hongli Yu, Xiaosheng Zhang, Xiangyue Wu, Chengdong Faculty of Robot Science and Engineering Northeastern University Shenyang China

Underwater image restoration is pivotal for enhancing the visual perception of underwater robots by improving image visibility and quality. Typically, underwater images are compromised by color distortion, low contrast, and noise, which degrade the robots' environmental perception. This paper introduces FS-Net, an innovative approach that synergistically utilizes information from both the frequency and spatial domains to overcome these challenges. By strategically separating and processing different frequency components, FS-Net effectively reduces noise and preserves intricate details. Concurrently, advanced spatial domain filtering and enhancement techniques are employed to augment contrast and rectify color distortions. This dual-domain integration ensures a comprehensive and efficient restoration process. We rigorously evaluate FS-Net using a well-established underwater image dataset, where it demonstrates substantial enhancements in both visual quality and quantitative performance metrics over existing methods. This work significantly advances underwater imaging technology, offering a robust solution that markedly improves image clarity and utility for underwater robotic applications. © Published under licence by IOP Publishing Ltd.

关键词： Underwater photography

来源：评论

学校读者我要写书评

暂无评论

Middle Code Prediction: Enhancing Code Generation for Uncommon Programming Languages in robotics 16

Middle Code Prediction: Enhancing Code Generation for Uncomm...

引用

16th Asian Conference on Machine Learning, ACML 2024

作者： Jia, Zixi Gao, Hongbin Li, Hexiao Faculty of Robot Science and Engineering Northeastern University Shenyang China

Generating executable code through natural language instructions to drive robotic movements is considered a crucial step towards achieving embodied intelligence. However, in the robotics domain, the scarcity of programming language data often necessitates manually encapsulating high-level APIs to enable Large Language Models(LLMs) to predict code correctly, which is time-consuming and incomplete. Therefore, this paper proposes a three-stage Middle Code Prediction(MCP) scheme, by injecting appropriate prompts at different stages, the LLMs can shift towards predicting middle code that it understands more easily. This middle code can then be converted into the final code through specific scripts, accomplishing the task of generating code in uncommon programming languages automatically and without the need for manually encapsulating high-level APIs. We tested our approach on Hospital Item Transport Dataset(HITD) and found that MCP could improve the mean accuracy of various baseline models to varying degrees, with an overall increase of 31%, while also enhancing the noise resistance of fine-tuned models. We conducted real-world experiments on industrial robotic arms, verifying the feasibility of MCP in scenarios with no API and partial API encapsulation. The method proposed in this paper provides a guideline for code generation in uncommon programming languages within the context of LLMs. Our experimental dataset is available at https://***/Ghbbbbb/MCP. © 2024 Z. Jia, H. Gao & H. Li.

关键词： robot programming

来源：评论

学校读者我要写书评

暂无评论

DCoT: Dual Chain-of-Thought Prompting for Large Multimodal Models 16

DCoT: Dual Chain-of-Thought Prompting for Large Multimodal M...

引用

16th Asian Conference on Machine Learning, ACML 2024

作者： Jia, Zixi Liu, Jiqiang Li, Hexiao Liu, Qinghua Gao, Hongbin Faculty of Robot Science and Engineering Northeastern University Shenyang China

Inference augmentation techniques such as Chain-of-Thought have already made their mark in Large Language Models (LLMs). However, transferring these advances to Large Multimodal Models (LMMs) presents greater challenges. Drawing inspiration from human cognitive processes, this paper proposes a plug-and-play Dual Chain-of-Thought strategy, a novel pipeline that combines visual and textual guidance to improve the performance of LMMs in complex multimodal tasks. The DCoT strategy uses a dual guidance mechanism to use bounding box markers to guide the model’s attention to the image region related to the query problem in the visual aspect, so as to achieve fine-grained image guidance, and in the text aspect, we propose a Fast In-Context Retrieval Framework (FICRF) dynamically and automatically obtains the most suitable examples from the well-built demonstration example cluster as context guidance according to the current problem. This bimodal approach that utilizes visual and textual guidance enhances the inference capabilities of LMMs. Extensive experiments on different LMMs and benchmark datasets have validated its effectiveness, opening up a new path in multimodal inference. Showcasing how the synergistic combination of visual and textual instructions can take the performance of these models to new heights, while demonstrating the potential of Chain-of-Thought and In-Context Learning as a superior alternative to the fine-tuning of LMMs. © 2024 Z. Jia, J. Liu, H. Li, Q. Liu & H. Gao.

关键词： In-Context Learning Large Multimodal Models Multimodal Chain-of-Thought

来源：评论

学校读者我要写书评

暂无评论

Attention-Based Generative Grasping Deformable Convolutional Neural Network 9

Attention-Based Generative Grasping Deformable Convolutional...

引用

9th International Conference on robotics and Automation engineering, ICRAE 2024

作者： Qin, Kang-Xiong Fu, Ru-Jie Zhang, Sheng-Wei Jiang, Yang Northeastern University Faculty of Robot Science and Engineering Shenyang China

ISBN: (纸本)9798331518301

Traditional generative grasping convolutional neural network suffers from insufficient feature selection capability and limited adaptability to complex geometric transformations in unstructured environments. To address these issues, this study proposes an optimizational method based on the Convolutional Block Attention Module (CBAM) and Deformable Convolutional Networks (DCN) to improve the performance of Generative Grasping Convolutional Neural Network (GG-CNN) in robotic grasping tasks. The Network is trained and evaluated using the Cornell and Jacquard datasets, the simulation and real-world experiments are conducted. The experimental results show that the AGG-DCNN exhibits better performance in model training and grasping tasks, achieving an accuracy of 97.19% on the Jacquard dataset. Additionally, the network shows significant improvements in convergence speed and robustness in learning accuracy. © 2024 IEEE.

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

Slot-Based Object-Centric Reinforcement Learning Algorithm 14

Slot-Based Object-Centric Reinforcement Learning Algorithm

引用

14th IEEE International Conference on Cyber Technology in Automation, Control, and Intelligent Systems, CYBER 2024

作者： Chen, Chao Wang, Fei Wang, Xinyao Northeastern University Faculty of Robot Science and Engineering China Northeastern University Faculty of Robot Science and Engineering Liaoning Shen Yang China Bussiness School Liaoning University China

ISBN: (纸本)9798331506056

With the development of reinforcement learning algorithms, it has become capable of handling a wide variety of complex tasks in simulated environments. However, applying reinforcement learning algorithms to real-world tasks still faces significant challenges. When facing vision-based tasks, these algorithms often embed images into a unified representation, which is effective for most simple tasks but inefficient when dealing with tasks involving interactions between multiple objects. This inefficiency arises from the lack of prior information about the world, neglecting the fundamental characteristic that the world is composed of multiple objects. This paper proposes an algorithm called SOCRL (Slot Based Object-Centric reinforcement learning), which enhances the efficiency of reinforcement learning algorithms in multi-object tasks by decoupling images into independent components. Experimental results demonstrate that our algorithm exhibits better generalization and higher sample efficiency compared to previous algorithms. © 2024 IEEE.

关键词： Adversarial machine learning

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：