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

Fully Asynchronous Neuromorphic Perception for Mobile robot Dodging With Loihi Chips

引用

IEEE Transactions on Automation science and engineering 2025年 22卷 12802-12815页

作者： Jiang, Junjie Kong, Delei Hu, Chenming Fang, Zheng Northeastern University Faculty of Robot Science and Engineering Shenyang110169 China Hunan University School of Robotics Changsha410082 China

Sparse and asynchronous sensing and processing in natural organisms lead to ultra low-latency and energy-efficient perception. Inspired by these characteristics, event cameras have attracted extensive attention from academia and industry. However, the mainstream event stream processing paradigms (e.g., event frames, 3D voxels) encounter issues such as feature loss, event stacking, and high computational burden. These problems deviate from the intended purpose of event cameras. To address these issues, we propose a fully asynchronous neuromorphic paradigm that integrates event cameras, spiking networks, and neuromorphic processors (Intel Loihi). This paradigm can faithfully process each event asynchronously as it arrives, mimicking the spike-driven signal processing in biological brains. We first propose a Key-Event-Point module to extract the key event stream from the raw event stream, effectively addressing the issue of limited transmission bandwidth when processing events on neuromorphic processors. Then, we propose the complete pipeline for the dodging network based on spiking neural networks to achieve offline training and online inference. Finally, we compare the proposed paradigm with event frames and 3D voxels processing paradigms in detail on the real mobile robot dodging task. Experimental results show that our scheme exhibits better robustness than image-like methods with different time windows and light conditions. Additionally, the energy consumption per inference of our scheme on the embedded Loihi processor is only 4.30% of that of the event spike tensor method on NVIDIA Jetson Orin NX with energy-saving mode, and 1.64% of that of the event frame method on the same neuromorphic processor. To the best of our knowledge, this is the first time that a fully asynchronous neuromorphic paradigm has been implemented for solving sequential tasks on a real mobile robot. © 2004-2012 IEEE.

关键词： Mobile robots

来源：评论

学校读者我要写书评

暂无评论

MCAFNet: Multiscale Channel Attention Fusion Network for Arbitrary Style Transfer

引用

IEEE Transactions on Instrumentation and Measurement 2025年 74卷

作者： Bai, Zhongyu Xu, Hongli Ding, Qichuan Zhang, Xiangyue Northeastern University Faculty of Robot Science and Engineering Shenyang110819 China

Recently, attention-based arbitrary style transfer (AST) techniques have been widely applied in image generation and video processing. However, the scale bias of the attention module used for contextual information extraction and multiscale feature aggregation poses a challenge in balancing the content structure and style patterns of images. In this work, a multiscale channel attention fusion network (MCAFNet) is proposed to generate stylization images with well-coordinated content and style. Specifically, the multiscale channel attention module (MCAM) is introduced to extract both local and global contextual information of style features within the channel dimension and subsequently aggregate this information with content features. Following MCAM, an attentional feature fusion module (AFFM) is adopted to effectively integrate both deep and shallow semantic features. Furthermore, a novel contrastive loss based on multi-source feature enhancement is proposed to optimize the spatial distribution between content and style features. Both qualitative and quantitative experimental results compared to the state-of-the-art (SOTA) baseline approaches indicate the superiority of the proposed method for real-time image and video style transfer. © 1963-2012 IEEE.

关键词： Information retrieval

来源：评论

学校读者我要写书评

暂无评论

CAO-RONet: A Robust 4D Radar Odometry with Exploring More Information from Low-Quality Points

arXiv

引用

arXiv 2025年

作者： Li, Zhiheng Cui, Yubo Huang, Ningyuan Pang, Chenglin Fang, Zheng Faculty of Robot Science and Engineering Northeastern University Shenyang China

Recently, 4D millimetre-wave radar exhibits more stable perception ability than LiDAR and camera under adverse conditions (e.g. rain and fog). However, low-quality radar points hinder its application, especially the odometry task that requires a dense and accurate matching. To fully explore the potential of 4D radar, we introduce a learning-based odometry framework, enabling robust ego-motion estimation from finite and uncertain geometry information. First, for sparse radar points, we propose a local completion to supplement missing structures and provide denser guideline for aligning two frames. Then, a context-aware association with a hierarchical structure flexibly matches points of different scales aided by feature similarity, and improves local matching consistency through correlation balancing. Finally, we present a window-based optimizer that uses historical priors to establish a coupling state estimation and correct errors of inter-frame matching. The superiority of our algorithm is confirmed on View-of-Delft dataset, achieving around a 50% performance improvement over previous approaches and delivering accuracy on par with LiDAR odometry. Our code will be available. Copyright © 2025, The Authors. All rights reserved.

关键词： Optical radar

来源：评论

学校读者我要写书评

暂无评论

Improving the Teaching Assessment System for the AI Comprehensive Practice Course Based on the K-means Algorithm within the CBA Framework 24

Improving the Teaching Assessment System for the AI Comprehe...

引用

3rd International Conference on Artificial Intelligence and Education, ICAIE 2024

作者： Xiao, Xibin Chu, Hao Wang, Fei Wen, Shiguang Faculty of Robot Science and Engineering Northeastern University Liaoning Shenyang China

ISBN: (纸本)9798400712692

Traditional assessment systems can't fully evaluate students' practical abilities in the course teaching process. To address the issue, we develop a teaching evaluation system for AI course. It is based on the Competition-Based Assessment (CBA) framework. The weights of evaluation criteria are optimized dynamically by using K-Means clustering and Spearman's rank correlation analysis. So that the system's accuracy and fairness has been enhanced. In the proposed system, the students' competencies are better reflected in the aspect of practical ability, innovation, and teamwork spirit. The experiment results show that our evaluation system provides a more objective assessment approach. Moreover, it can reflect students' practical skills effectively. © 2024 Copyright held by the owner/author(s). Publication rights licensed to ACM.

关键词： Assessment System CBA K-Means Spearman

来源：评论

学校读者我要写书评

暂无评论

Weakly Supervised Segmentation Framework for Thyroid Nodule Based on High-confidence Labels and High-rationality Losses

arXiv

引用

arXiv 2025年

作者： Chi, Jianning Li, Zelan Lin, Geng Sun, MingYang Yu, Xiaosheng Faculty of Robot Science and Engineering Northeastern University Shenyang110169 China

Weakly supervised segmentation methods can delineate thyroid nodules in ultrasound images efficiently using training data with coarse labels, but suffer from: 1) low-confidence pseudo-labels that simply follow topological priors, introducing significant label noise, and 2) low-rationality loss functions that rigidly compare segmentation with labels, ignoring discriminative information for nodules with diverse and complex shapes. To solve these problems, we clarify the objective together with required references of weakly supervised ultrasound image segmentation, and present the framework with high-confidence pseudo-labels to represent topological and anatomical information, and high-rationality losses to learn multi-level discriminative information. Specifically, we fuse geometric transformations of four-point annotations and results from the MedSAM model prompted by certain annotations to generate high-confidence box, foreground, and background labels. We design a high-rationality learning strategy comprising: 1) Alignment loss that measures the spatial projection consistency between the segmentation and box label, and the topological continuity of the segmentation within foreground label, guiding the network to perceive the location arrangement of nodule features;2) Contrastive loss that pulls features sampled from labeled foreground regions, while pushing features sampled from labeled foreground and background regions, guiding the network to capture the regional distribution of nodule and background features;3) Prototype correlation loss that measures the consistency between correlation maps derived by comparing features with foreground and background prototypes respectively, shrinking the uncertain regions to precise nodule edge delineation. Experimental results demonstrate that our method achieves state-of-the-art performance on the publicly available TN3K and DDTI datasets. The code is publicly available at HCL-HRL. © 2025, CC BY.

关键词： Ultrasonic imaging

来源：评论

学校读者我要写书评

暂无评论

Utilizing Large Language Models Enhanced by Chain-of-Thought for the Diagnosis of Typical Medical Cases 10th

Utilizing Large Language Models Enhanced by Chain-of-Though...

引用

10th China Health Information Processing Conference, CHIP 2024

作者： Liu, Jiqiang Liu, Chenyang Faculty of Robot Science and Engineering Northeastern University Shenyang China School of Computer Science and Engineering Northeastern University Shenyang China

ISBN: (纸本)9789819642977

In the past two years, the large language model has set off a new wave of research in the field of natural language processing, showing the ability of general-purpose artificial intelligence, which has been widely concerned by the industry. With the rapid development of medical vertical large model, its potential in clinical application has been paid more and more attention. However, there are gaps in the diagnosis of typical medical records. In order to continuously improve the application effect of medical large model in actual clinical scenarios and accelerate the implementation of medical large model, this paper describes the specific situation of our participation in the 10th China Conference on Health Information Processing (CHIP 2024). We used the improved LoRA method for fine tuning, and then used the Chain-of-Thought method for post-processing in the reasoning stage to improve performance. The experimental results show that the F1 score of our method in the final second round of evaluation reaches 0.9356, ranking second, which effectively verifies the generalization and robustness of our method. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.

关键词： Diagnosis

来源：评论

学校读者我要写书评

暂无评论

Intelligent interaction system based on multimodal conformal triboelectric nanogenerator patch for disabled sports and life

引用

science China(Technological sciences) 2025年第2期68卷 290-305页

作者： Yuzhang WEN Mengqi ZHANG Zhenning XIE Zida AN Bing LIU Fengxin SUN Tianming ZHAO Zuojun YU Fei WANG Yupeng MAO Physical Education Department Northeastern UniversityShenyang 110819China Faculty of Robot Science and Engineering Northeastern UniversityShenyang 110819China China Ice Sports College Beijing Sport UniversityBeijing 100084China State Key Laboratory of Robotics Shenyang Institute of AutomationChinese Academy of SciencesShenyang 110016China Criminal Investigation Police University of China Shenyang 110035China

Intelligent interaction technology based on advanced sensing units has become a research *** this paper,an intelligent interaction system(IIS)based on a multimodal conformal triboelectric nanogenerator patch(MCTP)is reported for sports monitoring and life promotion of the *** IIS consists of a cloud-based smart sports recognition program and a smart life *** porous micro-reinforced structure of polyurethane(PU)Sponge&Ecoflex,and the surface modification of PTFE by coating it with MXene endows the MCTP with excellent electrical *** shows good angle and force sensitivities of 1.1 V/°and 1.61 V/N,respectively,and has an ultra-fast response time of 15 ms,meeting the needs of human motion *** MCTP can also capture the mechanical energy,producing a maximum output power of 139μ*** with the commercial inertial measuring unit(IMU),the MCTP shows better consistency in human joint angle sensing.A precise analysis of wheelchair curling sports monitoring points is achieved by combining the MCTP and the human anatomy *** cloud-based smart sports recognition program utilizes machine learning to achieve visual feedback on the three throwing technologies of draw,takeout and *** addition,the developed smart life application uses a highsensitivity MCTP-sensor array design to achieve a variety of smart home interactions through simple finger ***,the IIS integrates the three major functions of smart sports sensing,energy capture,and smart home control,providing important application prospects in the fields of smart sports and human-computer interaction for the disabled.

关键词： triboelectric nanogenerator machine learning smart sports smart home human-computer interaction

来源：评论

学校读者我要写书评

暂无评论

Construction of an AI Course Cluster System for robotics Specialization Based on PBL Framework 24

Construction of an AI Course Cluster System for Robotics Spe...

引用

3rd International Conference on Artificial Intelligence and Education, ICAIE 2024

作者： Wang, Xiao Chu, Hao Ji, Peng Yu, Xiaosheng College Of Information Science and Engineering Northeastern University Liaoning Shenyang China Faculty of Robot Science and Engineering Northeastern University Liaoning Shenyang China

ISBN: (纸本)9798400712692

Traditional AI-related courses in robotics major are not related closely to the latest technological advancements. In order to explore a better approach in the teaching system, this paper introduces a new AI course cluster system. The system conducts the PBL teaching structure and designs a tailored AI course cluster for robotics students. An organized and project-based teaching model is established too. We also conduct case studies to the students in the robotics major. The experimental results verify the effective of the proposed cluster system. © 2024 Copyright held by the owner/author(s). Publication rights licensed to ACM.

关键词： AI courses cluster system PBL robotics major

来源：评论

学校读者我要写书评

暂无评论

Hierarchical Reinforcement Learning for Safe Mapless Navigation with Congestion Estimation

arXiv

引用

arXiv 2025年

作者： Gao, Jianqi Pang, Xizheng Liu, Qi Li, Yanjie The Department of Guangdong Key Laboratory of Intelligent Morphing Mechanisms and Adaptive Robotics School of Inteligence Science and Engineering Harbin Institute of Technology Shenzhen518055 China The Faculty of Robot Science and Engineering Northeastern University Shenyang110819 China

Reinforcement learning-based mapless navigation holds significant potential. However, it faces challenges in indoor environments with local minima area. This paper introduces a safe mapless navigation framework utilizing hierarchical reinforcement learning (HRL) to enhance navigation through such areas. The high-level policy creates a sub-goal to direct the navigation process. Notably, we have developed a sub-goal update mechanism that considers environment congestion, efficiently avoiding the entrapment of the robot in local minimum areas. The low-level motion planning policy, trained through safe reinforcement learning, outputs real-time control instructions based on acquired sub-goal. Specifically, to enhance the robot’s environmental perception, we introduce a new obstacle encoding method that evaluates the impact of obstacles on the robot’s motion planning. To validate the performance of our HRL-based navigation framework, we conduct simulations in office, home, and restaurant environments. The findings demonstrate that our HRL-based navigation framework excels in both static and dynamic scenarios. Finally, we implement the HRL-based navigation framework on a TurtleBot3 robot for physical validation experiments, which exhibits its strong generalization capabilities. Copyright © 2025, The Authors. All rights reserved.

关键词： robot programming

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：