检索结果-内蒙古大学图书馆

arXiv 2025年

作者： Liu, Bangzhen Zheng, Chenxi Xu, Xuemiao Xu, Cheng Zhang, Huaidong He, Shengfeng South China University of Technology Guangzhou China Guangdong Engineering Center for Large Model and GenAI Technology The State Key Laboratory of Subtropical Building and Urban Science The Ministry of Education Key Laboratory of Big Data and Intelligent Robot The Guangdong Provincial Key Lab of Computational Intelligence and Cyberspace Information China Singapore Management University Singapore

The vulnerability of 3D point cloud analysis to unpredictable rotations poses an open yet challenging problem: orientation-aware 3D domain generalization. Cross-domain robustness and adaptability of 3D representations are crucial but not easily achieved through rotation augmentation. Motivated by the inherent advantages of intricate orientations in enhancing generalizability, we propose an innovative rotation-adaptive domain generalization framework for 3D point cloud analysis. Our approach aims to alleviate orientational shifts by leveraging intricate samples in an iterative learning process. Specifically, we identify the most challenging rotation for each point cloud and construct an intricate orientation set by optimizing intricate orientations. Subsequently, we employ an orientation-aware contrastive learning framework that incorporates an orientation consistency loss and a margin separation loss, enabling effective learning of categorically discriminative and generalizable features with rotation consistency. Extensive experiments and ablations conducted on 3D cross-domain benchmarks firmly establish the state-of-the-art performance of our proposed approach in the context of orientation-aware 3D domain generalization. © 2025, CC BY-NC-SA.

关键词： Contrastive Learning

来源：评论

学校读者我要写书评

暂无评论

Implicit Coarse-to-Fine 3D Perception for Category-level Object Pose Estimation from Monocular RGB Image

Implicit Coarse-to-Fine 3D Perception for Category-level Obj...

引用

IEEE International Conference on robotics and Automation (ICRA)

作者： Jia Li Li Jin Xibin Song Yeheng Chen Nan Li Xueying Qin School of Software Shandong University P.R. China Engineering Research Center of Digital Media Technology Ministry of Education P.R. China XR Vision Labs Tencent P.R. China Intelligent Robot Research Center Zhejiang Lab P.R. China

ISBN: (数字)9798350384574

ISBN: (纸本)9798350384581

Category-level object pose estimation demonstrates robust generalization capabilities that benefit robotics applications. However, exclusive reliance on RGB images without leveraging any 3D information introduces ambiguity in the translation and size of objects, leading to suboptimal performance. In this paper, we propose a framework for category-level pose estimation from a single RGB image in an end-to-end manner, i.e., Feature Auxiliary Perception Network (FAP-Net). To address inaccurate pose estimation caused by the inherent ambiguity of RGB images, we design a coarse-to-fine approach that first harnesses geometry supervision to facilitate coarse 3D feature perception and subsequently refines the features based on pose and size constraints. Experimental results on REAL275 and CAMERA25 demonstrate that FAP-Net achieves significant improvements (14.7% on 10°10cm and 11.4% on IoU50 on the real-scene REAL275 dataset) over the state-of-the-art and real-time inference (42 FPS).

关键词： Geometry Three-dimensional displays Pose estimation Grasping Robustness Real-time systems robots

来源：评论

学校读者我要写书评

暂无评论

Unveiling the Complete Variant of Spherical robots

arXiv

引用

arXiv 2024年

作者： Nigatu, Hassen Jihao, Li Shi, Gaokun Lu, Guodong Dong, Huixu Robot Perception and Grasp Lab School of Mechanical Engineering Zhejiang University Hangzhou310027 China Yuyao Technology Innovation Center No.479 Ye Shan Road Yuyao Zhejiang Province Ningbo Shi China The State Key Laboratory of Fluid Power and Mechatronic Systems School of Mechanical Engineering The Engineering Research Center for Design Engineering and Digital Twin of Zhejiang Province School of Mechanical Engineering Zhejiang University Hangzhou310027 China

This study presents a systematic enumeration of spherical (SO(3)) type parallel robots’ variants using an analytical velocity-level approach. These robots are known for their ability to perform arbitrary rotations around a fixed point, making them suitable for numerous applications. Despite their architectural diversity, existing research has predominantly approached them on a case-by-case basis. This approach hinders the exploration of all possible variants, thereby limiting the benefits derived from architectural diversity. By employing a generalized analytical approach through the reciprocal screw method, we systematically explore all the kinematic conditions for limbs yielding SO(3) ***, all 73 possible types of non-redundant limbs suitable for generating the target SO(3) motion are identified. The approach involves performing an in-depth algebraic motion-constraint analysis and identifying common characteristics among different variants. This leads us to systematically explore all 73 symmetric and 5256 asymmetric variants, which in turn become a total of 5329, each potentially having different workspace capability, stiffness performance, and dynamics. Hence, having all these variants can facilitate the innovation of novel spherical robots and help us easily find the best and optimal ones for our specific applications. © 2024, CC BY.

关键词： robots

来源：评论

学校读者我要写书评

暂无评论

Shared Neuronal Bases of Inhibition And Economic Choice in Orbitofrontal Cortex

SSRN

引用

SSRN 2023年

作者： Balasubramani, Pragathi Priyadharsini Yadav, Peeyushi Hayden, Benjamin Y. Translational Neuroscience and Technology Lab Department of Cognitive Science The Mehta Family Center for Engineering and Medicine IIT Kanpur Indian Institute of Technology Kanpur Uttar Pradesh Kanpur208016 India Department of Neurosurgery Baylor College of Medicine HoustonTX77401 United States

Economic choice and inhibition are two important elements of our cognitive repertoires that may reflect shared computations. During economic choice, options are typically considered serially;this asynchronous contemplation means that each option is subject to an accept-reject decision;such decisions may have some kinship with stopping decisions. One prediction of this idea is that stopping and choice may rely on similar neural processes occurring in overlapping brain circuits. To explore this idea, we recorded neuronal activity in orbitofrontal cortex (OFC) Area 13 while macaques performed a stop signal task interleaved with a structurally similar choice task. We were able to distinguish between stopping and choice with significant task-related neural components using demixed principal component analysis. However, using neural network decoders, we find that OFC ensembles have overlapping codes for stopping and choice: the decoder that was only trained to identify accept vs. reject trials performed with higher efficiency even when tested on the stop trials. These results provide tentative support for the idea that mechanisms underlying inhibitory control and choice selection may be subject to theoretical unification. © 2023, The Authors. All rights reserved.

关键词： Principal component analysis

来源：评论

学校读者我要写书评

暂无评论

ConvBLS: An Effective and Efficient Incremental Convolutional Broad Learning System for Image Classification

arXiv

引用

arXiv 2023年

作者： Lei, Chunyu Chen, C.L. Philip Guo, Jifeng Zhang, Tong The Guangdong Provincial Key Laboratory of Computational Intelligence and Cyberspace Information The School of Computer Science and Engineering South China University of Technology Guangzhou510006 China The Brain and Affective Cognitive Research Center Pazhou Lab Guangzhou510335 China

Deep learning generally suffers from enormous computational resources and time-consuming training processes. Broad Learning System (BLS) and its convolutional variants have been proposed to mitigate these issues and have achieved superb performance in image classification. However, the existing convolutional-based broad learning system (C-BLS) either lacks an efficient training method and incremental learning capability or suffers from poor performance. To this end, we propose a convolutional broad learning system (ConvBLS) based on the spherical K-means (SKM) algorithm and two-stage multi-scale (TSMS) feature fusion, which consists of the convolutional feature (CF) layer, convolutional enhancement (CE) layer, TSMS feature fusion layer, and output layer. First, unlike the current C-BLS, the simple yet efficient SKM algorithm is utilized to learn the weights of CF layers. Compared with random filters, the SKM algorithm makes the CF layer learn more comprehensive spatial features. Second, similar to the vanilla BLS, CE layers are established to expand the feature space. Third, the TSMS feature fusion layer is proposed to extract more effective multi-scale features through the integration of CF layers and CE layers. Thanks to the above design and the pseudo-inverse calculation of the output layer weights, our proposed ConvBLS method is unprecedentedly efficient and effective. Finally, the corresponding incremental learning algorithms are presented for rapid remodeling if the model deems to expand. Experiments and comparisons demonstrate the superiority of our method. Copyright © 2023, The Authors. All rights reserved.

关键词： Learning systems

来源：评论

学校读者我要写书评

暂无评论

A Novel Soft Actuator Based on Mini DC Motor and Supercoiled Polymer Artificial Muscle

A Novel Soft Actuator Based on Mini DC Motor and Supercoiled...

引用

2022 IEEE International Conference on Real-Time Computing and robotics, RCAR 2022

作者： Zhu, Honghui Jin, Kaixiang Yang, Yang Wei, Zhong Liu, Jia Kang, Long Li, Yunquan Nanjing University of Information Science and Technology School of Automation Nanjing210044 China Jiangsu Province Engineering Research Center of Intelligent Meteorological Exploration Robot China Nanjing University of Science and Technology PCA Lab Key Lab. of Intelligent Percept. and Syst. for High-Dimensional Info. of Min. of Education Jiangsu Key Lab of Image and Video Understanding for Social Secty Sch. of Comp. Sci. and Eng Nanjing210094 China South China University of Technology Panyu District Shien-Ming Wu School of Intelligent Engineering Guangzhou China

ISBN: (数字)9781665469838

ISBN: (纸本)9781665469838

This paper describes a novel design of a soft actuator that combines two types of actuation methods: DC motor actuation and supercoiled polymer (SCP) artificial muscle. SCP artificial muscle is used to substitute the traditional cable and attached to the gearmotor. The hybrid actuator intends to alleviate the drawbacks of both actuation methods. This novel actuator features a fast actuation speed and the addition of SCP artificial muscles endow it with a certain degree of compliance. At the same time, the thermal contraction property of SCP artificial muscle can also increase its output force. A soft robotic finger is designed to conFigure the actuator in it. In the experiment section, finger's stiffness and output force tests are performed under three different actuation methods (single SCP/gearmotor kite line/gearmotor SCP). Experiment results demonstrate the superiority of this novel actuator in terms of compliance (compared with gearmotor kite line actuation), actuation speed, and output force (compared with single SCP actuation). This actuator has the potential to be used for grasping and manipulation of soft robots in the future. © 2022 IEEE.

关键词： Actuators

来源：评论

学校读者我要写书评

暂无评论

The Stiffness of 3-PRS PM Across Parasitic and Orientational Workspace

arXiv

引用

arXiv 2024年

作者： Nigatu, Hassen Jihao, Li Zhu, Keqi Zhang, Junhan Guo, Haotian Lu, Guodong Kim, Doik Robot Perception and Grasp Lab School of Mechanical Engineering Zhejiang University Hangzhou310027 China Robotics Research Center of Yuyao Robotics Institute Zhejiang University Yuyao Technology Innovation Center No.479 Ye Shan Road Yuyao Zhejiang Province Ningbo China The State Key Laboratory of Fluid Power and Mechatronic Systems School of Mechanical Engineering The Engineering Research Center for Design Engineering and Digital Twin of Zhejiang Province Zhejiang University Hangzhou310027 China Korea Republic of

This study investigates the stiffness characteristics of the Sprint Z3 head, also known as 3-PRS Parallel Kinematics Machines, which are among the most extensively researched and viably successful manipulators for precision machining applications. Despite the wealth of research on these robotic manipulators, no previous work has demonstrated their stiffness performance within the parasitic motion space. Such an undesired motion influences their stiffness properties, as stiffness is configuration-dependent. Addressing this gap, this paper develops a stiffness model that accounts for both the velocity-level parasitic motion space and the regular workspace. Numerical simulations are provided to illustrate the stiffness characteristics of the manipulator across all considered spaces. The results indicate that the stiffness profile within the parasitic motion space is both shallower and the values are smaller when compared to the stiffness distribution across the orientation workspace. This implies that evaluating a manipulator’s performance adequately requires assessing its ability to resist external loads during parasitic motion. Therefore, comprehending this aspect is crucial for redesigning components to enhance overall stiffness. © 2024, CC BY.

关键词： Manipulators

来源：评论

学校读者我要写书评

暂无评论

FR²Seg: Continual Segmentation Across Multiple Sites via Fourier Style Replay and Adaptive Consistency Regularization 39

FR<sup>2</sup>Seg: Continual Segmentation Across Multiple Si...

引用

39th Annual AAAI Conference on Artificial Intelligence, AAAI 2025

作者： Xu, Cheng Zhang, Weiwen Zhang, Hongrui Xu, Xuemiao Zhang, Huaidong Zou, Jing Qin, Jing South China University of Technology China The Hong Kong Polytechnic University Hong Kong Guangdong Engineering Center for Large Model and GenAI Technology China State Key Laboratory of Subtropical Building Science China Ministry of Education Key Laboratory of Big Data and Intelligent Robot China Guangdong Provincial Key Lab of Computational Intelligence and Cyberspace Information China

ISBN: (纸本)157735897X

In clinical imaging, medical segmentation networks typically require continually adapting to new data from multiple sites over time, as aggregating all data for learning at once can be impractical due to storage limitations and privacy concerns. However, existing methods basically overlook domain-specific characteristics and fall short of adequately capturing domain-invariant knowledge during continual learning, leading to undesired catastrophic forgetting of previous sites and inferior generalization to new sites. To tackle this issue, this paper introduces FR²Seg, to sufficiently exploit both domain-specific and domain-invariant knowledge for efficient continual learning with the aid of low-frequency cues. For the former aspect, we propose a Fourier style replay module to synthesize pseudo images with old-site styles for data augmentation during new-site training, effectively preventing catastrophic forgetting without sacrificing data privacy. For the latter, we present a Fourier adaptive consistency regularization to identify and constrain the optimization of domain-invariant parameters with explicit awareness of knowledge transferability across sites, ensuring excellent generalizability to new sites. Experimental results on two public datasets confirm our method's superiority over existing state-of-the-art continual learning methods. Copyright © 2025, Association for the Advancement of Artificial Intelligence (***). All rights reserved.

关键词： Image segmentation

来源：评论

学校读者我要写书评

暂无评论

Kinematic-Model-Free Tip Position Control of Reconfigurable and Growing Soft Continuum robots

Kinematic-Model-Free Tip Position Control of Reconfigurable ...

引用

IEEE International Conference on Soft robotics (RoboSoft)

作者： Ahmad AlAttar Ikhlas Ben Hmida Federico Renda Petar Kormushev Robot Intelligence Lab Dyson School of Design Engineering Imperial College London UK Robotics Lab Dubai Future Labs Dubai UAE Department of Mechanical Engineering Khalifa University of Science and Technology Abu Dhabi UAE Khalifa University Center for Autonomous Robotic Systems (KUCARS) Khalifa University Abu Dhabi UAE

Soft robots have many advantages over their rigid counterparts. These include their inherent compliance, lightweight and high adaptability to cluttered workspaces. Soft continuum robots, biologically inspired snake-like robots, are hyper-redundant and highly deformable. These robots can be challenging to control due to their complex kinematic and dynamic models. This paper presents a novel kinematic-model-free controller that uses a quasi-static assumption in order to control the tip-position of soft continuum robots with threadlike actuation while compensating for gravity simultaneously. The controller was tested on simulated continuum soft robots to demonstrate its ability to guide the tip while following a given trajectory. Novel kinematic-model-free control methods are introduced for soft robots' route and length control. The robustness of the controller is demonstrated with an actuator-failure test. The kinematic-model-free controller provides an adaptive control method for static, re-configuring, and growing soft continuum robots with threadlike actuation.

关键词： Actuators Adaptation models Three-dimensional displays Biological system modeling Position control Kinematics Soft robotics

来源：评论

学校读者我要写书评

暂无评论

Detection-and-Excitation Neural Network Achieves Accurate Nasopharyngeal Carcinoma Segmentation in Multi-modality MR Images

Detection-and-Excitation Neural Network Achieves Accurate Na...

引用

2021 IEEE International Conference on Bioinformatics and Biomedicine, BIBM 2021

作者： Tao, Guihua Li, Haojiang Liu, Lizhi Cai, Hongmin South China University of Technology School of Computer Science and Engineering Guangzhou China Sun Yat-sen University Cancer Center Key Laboratory of Nasopharyngeal Carcinoma Diagnosis and Therapy Collaborative Innovation Center for Cancer Medicine Guangzhou China Brain and Affective Cognitive Research Center Pazhou Lab Guangzhou China

ISBN: (纸本)9781665401265

Accurate and reliable segmentation of nasopharyngeal carcinoma (NPC) in magnetic resonance image (MRI) is important for treatment planning and follow-up evaluation. However, it is still challenging because NPC is infiltrative with a vague border and has a tiny volume with varying sizes and shapes. The above problems easily cause the NPC tumors' features to submerge in the process of feature extraction. Standard segmentation methods do not cope with the 'feature submergence' and performed unsatisfactorily in NPC segmentation. To address the problem, a dual-supervised method equipped with the detection-and-excitation module (DEM) was proposed. DEM strengthens the region of interest (ROI) and weakens complex and immense background through incorporating detection feature maps (outline maps) with intermediated segmentation features. To be specific, the DEM explored the outline of the ROI and output a detection feature map by using a supervised layer guided by a tailored loss function. Then, the DEM uses the results to recalibrates the intermediated segmentation. Finally, a neural network equipped with the proposed DEM was elaborately designed to achieve accurate segmentation. We applied the proposed network termed as DENet on a real nasopharyngeal carcinoma dataset to realize automatic NPC tumors segmentation in multi-modality MR images. The proposed DEM enables the neural network to segment NPC tumors within the immense and complex background. Experimental results have demonstrated the effectiveness of the proposed method by comparing it with benchmark models. © 2021 IEEE.

关键词： Tumors

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：