检索结果-内蒙古大学图书馆

arXiv 2025年

作者： Chen, Zizhi Han, Minghao Zhang, Xukun Ma, Shuwei Liu, Tao Wei, Xing Zhang, Lihua Academy for Engineering and Technology Fudan University Shanghai China Institute of Metaverse & Intelligent Medicine Fudan University Shanghai China Engineering Research Center of AI and Robotics Ministry of Education Shanghai China Jilin Provincial Key Laboratory of Intelligence Science and Engineering Changchun China

Multimodal learning combining pathology images and genomic sequences enhances cancer survival analysis but faces clinical implementation barriers due to limited access to genomic sequencing in under-resourced regions. To enable survival prediction using only whole-slide images (WSI), we propose the Visual-Genomic Answering-Guided Transformer (VGAT), a framework integrating Visual Question Answering (VQA) techniques for genomic modality reconstruction. By adapting VQA’s text feature extraction approach, we derive stable genomic representations that circumvent dimensionality challenges in raw genomic data. Simultaneously, a cluster-based visual prompt module selectively enhances discriminative WSI patches, addressing noise from unfiltered image regions. Evaluated across five TCGA datasets, VGAT outperforms existing WSI-only methods, demonstrating the viability of genomic-informed inference without sequencing. This approach bridges multimodal research and clinical feasibility in resource-constrained settings. The code link is https://***/CZZZZZZZZZZZZZZZZZ/VGAT. Copyright © 2025, The Authors. All rights reserved.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Vector Field-Guided Circumnavigation for an Unmanned Surface Vehicle With Boundary and Obstacle Constraints

引用

IEEE Transactions on Emerging Topics in Computing 2025年

作者： Wu, Nailong Wang, Jigang Wang, Yueying He, Shuping Donghua University College of Information Science and Technology Shanghai201620 China Donghua University Engineering Research Center of Digitized Textile & Apparel Technology Ministry of Education Shanghai201620 China Shanghai University School of Mechatronic Engineering and Automation Shanghai200444 China Anhui University Key Laboratory of Intelligent Computing and Signal Processing Ministry of Education School of Electrical Engineering and Automation Hefei230601 China

This article aims to tackle the challenge of rapid target circumnavigation by unmanned surface vehicles (USVs) under boundary and obstacle constraints through leveraging an emerging computing architecture specifically tailored for unmanned systems. Owing to the underactuation of USV dynamic systems and the presence of various time-varying surface disturbances, attaining rapid and accurate position control along with maneuverability poses a significant challenge to USVs. Therefore, when an USV maneuvers around a target in a confined search area with obstacles, it is essential for the guidance law to consider collision risks and environmental disturbances. This paper proposes a modified Lyapunov vector field with heading deviation compensation to circle the target in complex search regions. The vector field is designed to increase the radial velocity of the USV and accelerate convergence towards the target circle during the pursuit phase. By combining the proposed Lyapunov vector field with a boundary vector field, the USV can swiftly avoid boundaries and dynamic obstacles and then continue circling the target. The proposed computing architecture extends the application scope of the target circumnavigation to practical water surface environments, which is validated by extensive simulations and rigorous experiments. © 2025 IEEE.

关键词： Unmanned surface vehicles

来源：评论

学校读者我要写书评

暂无评论

Influence of laser powder bed fusion of high-entropy alloy transition layer on the wetting and spreading behaviour of Al alloy on steel substrate surface

引用

Journal of Materials Processing technology 2025年 340卷

作者： Zeng, Yijian Yang, Jin Dou, Tianyu Zheng, Min Zhao, Yixuan Oliveira, J.P. Shen, Jiajia Tan, Caiwang Xia, Hongbo Zhang, Hua School of Materials Science and Engineering Shanghai University of Engineering Science Shanghai201620 China School of Materials Science and Engineering Nanchang Hangkong University Nanchang330063 China CENIMAT|I3N Department of Materials Science School of Science and Technology NOVA University Lisbon Caparica Portugal State Key Laboratory of Precision Welding & Joining of Materials and Structures Harbin Institute of Technology Harbin150001 China Shanghai Collaborative Innovation Center of Intelligent Manufacturing Robot Technology for Large Components Shanghai201620 China

The high-quality joining of Al alloys to steel is widely used in automotive manufacturing. However, because of the significant differences in their thermophysical properties, the wetting of Al alloys on steel is very limited, which affects the joining quality. To address this challenge, the laser powder bed fusion (LPBF) technique was used to prepare an FeCoNiCrMn high entropy alloy (HEA) with a dense transition layer (DTL) and micropillar transition layer (MTL) on a steel substrate for the first time. Then, the dynamic wetting and spreading behaviours of Al-12Si alloy over the transition layer-free, DTL, and MTL steels were comparatively studied. The spreading mechanisms were revealed by the spreading dynamics and an analysis of the interfacial microstructure. This study investigated a new method to improve the wetting of Al alloys on steel substrates and elucidated the mechanism of the effect of the HEA transition layer formed using LPBF on Al/steel wetting, which may provide some guidance for improving brazing, welding, coating, and other processes involving solid/liquid interfacial interactions. © 2025

关键词： Aluminum alloys

来源：评论

学校读者我要写书评

暂无评论

Learning to Learn Transferable Generative Attack for Person Re-Identification

引用

IEEE Transactions on Image Processing 2025年 PP卷 PP页

作者： Bian, Yuan Liu, Min Wang, Xueping Ma, Yunfeng Wang, Yaonan Hunan University National Engineering Research Center of Robot Visual Perception and Control Technology College of Electrical and Information Engineering Hunan Changsha China Hunan Normal University Hunan Provincial Key Laboratory of Intelligent Computing and Language Information Processing College of Information Science and Engineering Hunan Changsha China

Deep learning-based person re-identification (reid) models are widely employed in surveillance systems and inevitably inherit the vulnerability of deep networks to adversarial attacks. Existing attacks merely consider cross-dataset and cross-model transferability, ignoring the cross-test capability to perturb models trained in different domains. To powerfully examine the robustness of real-world re-id models, the Meta Transferable Generative Attack (MTGA) method is proposed, which adopts meta-learning optimization to promote the generative attacker producing highly transferable adversarial examples by learning comprehensively simulated transfer-based crossmodel&dataset&test black-box meta attack tasks. Specifically, cross-model&dataset black-box attack tasks are first mimicked by selecting different re-id models and datasets for meta-train and meta-test attack processes. As different models may focus on different feature regions, the Perturbation Random Erasing module is further devised to prevent the attacker from learning to only corrupt model-specific features. To boost the attacker learning to possess cross-test transferability, the Normalization Mix strategy is introduced to imitate diverse feature embedding spaces by mixing multi-domain statistics of target models. Extensive experiments show the superiority of MTGA, especially in cross-model&dataset and cross-model&dataset&test attacks, our MTGA outperforms the SOTA methods by 20.0% and 11.3% on mean mAP drop rate, respectively. © 1992-2012 IEEE.

关键词： Generative adversarial networks

来源：评论

学校读者我要写书评

暂无评论

Hierarchy-Boosted Funnel Learning for Identifying Semiconductors with Ultralow Lattice Thermal Conductivity

arXiv

引用

arXiv 2025年

作者： Wu, Mengfan Yan, Shenshen Ren, Jie Shanghai Research Institute for Intelligent Autonomous Systems Tongji University Shanghai200092 China Center for Phononics and Thermal Energy Science Shanghai Key Laboratory of Special Artificial Microstructure Materials and Technology School of Physics Science and Engineering Tongji University Shanghai200092 China

Data-driven machine learning (ML) has demonstrated tremendous potential in material property predictions. However, the scarcity of materials data with costly property labels in the vast chemical space presents a significant challenge for ML in efficiently predicting properties and uncovering structure-property relationships. Here, we propose a novel hierarchy-boosted funnel learning (HiBoFL) framework, which is successfully applied to identify semiconductors with ultralow lattice thermal conductivity (κL). By training on only a few hundred materials targeted by unsupervised learning from a pool of hundreds of thousands, we achieve efficient and interpretable supervised predictions of ultralow κL, thereby circumventing large-scale brute-force ab initio calculations without clear objectives. As a result, we provide a list of candidates with ultralow κL for potential thermoelectric applications and discover a new factor that significantly influences structural anharmonicity. This HiBoFL framework offers a novel practical pathway for accelerating the discovery of functional materials. © 2025, CC BY-NC-ND.

关键词： Structural properties

来源：评论

学校读者我要写书评

暂无评论

Vehicular Road Crack Detection with Deep Learning: A New Online Benchmark for Comprehensive Evaluation of Existing Algorithms

arXiv

引用

arXiv 2025年

作者： Ma, Nachuan Song, Zhengfei Hu, Qiang Liu, Chuang-Wei Han, Yu Zhang, Yanting Fan, Rui Xie, Lihua College of Electronics & Information Engineering Shanghai Research Institute for Intelligent Autonomous Systems the State Key Laboratory of Intelligent Autonomous Systems Frontiers Science Center for Intelligent Autonomous Systems Tongji University Shanghai201804 China School of Computer Science and Technology Donghua University Shanghai201620 China School of Electrical and Electronic Engineering Nanyang Technological University 50 Nanyang Avenue Singapore639798 Singapore

In the emerging field of urban digital twins (UDTs), advancing intelligent road inspection (IRI) vehicles with automatic road crack detection systems is essential for maintaining civil infrastructure. Over the past decade, deep learning-based road crack detection methods have been developed to detect cracks more efficiently, accurately, and objectively, with the goal of replacing manual visual inspection. Nonetheless, there is a lack of systematic reviews on state-of-the-art (SoTA) deep learning techniques, especially data-fusion and label-efficient algorithms for this task. This paper thoroughly reviews the SoTA deep learning-based algorithms, including (1) supervised, (2) unsupervised, (3) semi-supervised, and (4) weakly-supervised methods developed for road crack detection. Also, we create a dataset called UDTIRI-Crack, comprising 2, 500 high-quality images from seven public annotated sources, as the first extensive online benchmark in this field. Comprehensive experiments are conducted to compare the detection performance, computational efficiency, and generalizability of public SoTA deep learning-based algorithms for road crack detection. In addition, the feasibility of foundation models and large language models (LLMs) for road crack detection is explored. Afterwards, the existing challenges and future development trends of deep learning-based road crack detection algorithms are discussed. We believe this review can serve as practical guidance for developing intelligent road detection vehicles with the next-generation road condition assessment systems. The released benchmark UDTIRI-Crack is available at https://***/submission/. © 2025, CC BY.

关键词： Self-supervised learning

来源：评论

学校读者我要写书评

暂无评论

Two-Stage Hand-Eye Calibration based on Variance Minimization Principle

引用

IEEE Transactions on Instrumentation and Measurement 2025年 74卷

作者： Ding, Tao Tian, Dazhuang Wu, Hao Zhu, Dahu Li, Wenlog Wuhan University of Technology Hubei Key Laboratory of Advanced Technology for Automotive Components School of Automotive Engineering Wuhan430070 China Wuhan University of Technology Hubei Digital Manufacturing Key Laboratory School of Mechanical and Electronic Engineering Wuhan430070 China Wuhan University of Technology Hubei Key Laboratory of Advanced Technology for Automotive Components Hubei Provincial Engineering Research Center of Robot and Intelligent Manufacturing Wuhan430070 China Huazhong University of Science and Technology State Key Laboratory of Intelligent Manufacturing Equipment and Technology Wuhan430074 China

Hand-eye calibration in industrial scenarios greatly determines the efficiency and effectiveness of 3D vision-guided robotic tasks. Most of the existing hand-eye calibration methods, however, rely on collecting large-scale point cloud samples to guarantee high measurement accuracy, which is time-consuming and particularly influenced by measurement defects such as local missing, noise, and uneven density. To overcome this challenging problem, we present a novel two-stage hand-eye calibration method based on the variance minimization principle. In the first stage, a robust function weighted variance minimization (RFWVM) registration algorithm is proposed for fine registration of defective point clouds, aiming at solving the coarse hand-eye calibration matrix with few-shot point clouds. In the second stage, an objective function of the calibration error correction algorithm is established via the point-to-plane distance variance minimization, which is intended to suppress the influence of measurement defects. Through these two stages, the globally optimal hand-eye matrix is solved by iteratively correcting the coarse hand-eye calibration matrix. A series of experiments validate the universality, accuracy, and robustness of the proposed method. © 1963-2012 IEEE.

关键词： Error correction

来源：评论

学校读者我要写书评

暂无评论

Decoupling of nanoindentation residual stress field in single-crystalline 4H-SiC via micro-Raman spectroscopy

引用

Optics Letters 2025年第11期50卷 3513-3516页

作者： Yang, Zhoudong Tian, Jing Wang, Xinyue Chen, Junwei Zuo, Yuanhui Zhang, Rongjun Tang, Hongyu Fan, Xuejun Zhang, Guoqi Fan, Jiajie Shanghai Engineering Technology Research Center for SiC Power Device School of Intelligent Robotics and Advanced Manufacturing Fudan University Shanghai200433 China Research Institute of Fudan University in Ningbo Ningbo315336 China School of Information Science and Technology Fudan University Shanghai200433 China Department of Mechanical Engineering Lamar University PO Box 10028 BeaumontTX77710 United States EEMCS Faculty Delft University of Technology Delft2628 Netherlands

This Letter presents a combined analytical and experimental method to effectively decouple the radial and tangential residual stress fields induced by Berkovich nanoindentation in single-crystalline 4H-SiC using micro-Raman spectroscopy. By integrating the Raman stress characterization model with Yoffe’s expanding cavity model, precise extraction of individual residual stress components around the indentation region is realized. Through the vertical backscattering micro-Raman mapping of the E2 phonon mode, we systematically investigate the residual stress distribution near the indentation. The results highlight significant anisotropy in nanoindentation-induced stress fields, strongly dependent on the crystal orientation of 4H-SiC, predominantly featuring radial tensile stress gradients. This comprehensive theoretical–experimental approach offers a robust optical framework for residual stress characterization in 4H-SiC and provides foundational insights for extending Raman spectroscopy-based stress characterization to other crystalline materials and related device structures. © 2025 Optica Publishing Group. All rights, including for text and data mining (TDM), Artificial Intelligence (AI) training, and similar technologies, are reserved.

关键词： Raman spectroscopy

来源：评论

学校读者我要写书评

暂无评论

Catalytic remodeling of complex alkenes to oxonitriles through C=C double bond deconstruction

引用

Science (New York, N.Y.) 2025年第6738期387卷 1083-1090页

作者： Cheng, Zengrui Huang, Kaimeng Wang, Chen Chen, Lili Li, Xinyao Hu, Zhibin Shan, Xinyuan Cao, Peng-Fei Sun, Haofeng Chen, Wei Li, Chenhao Zhang, Ziyao Tan, Hui Jiang, Xue Zhang, Guikai Zhang, Zhongying Lin, Min Wang, Liang Zheng, Anmin Xia, Changjiu Wang, Teng Song, Song Shu, Xingtian Jiao, Ning State Key Laboratory of Natural and Biomimetic Drugs Chemical Biology Center School of Pharmaceutical Sciences Peking University Beijing China State Key Laboratory of Catalytic Materials and Reaction Engineering Research Institute of Petroleum Processing SINOPEC Beijing China Department of Chemistry College of Sciences Shanghai Engineering Research Center of Organ Repair Shanghai University Shanghai China College of Chemistry State Key Laboratory of Organic-Inorganic Composites Beijing University of Chemical Technology Beijing China Key Laboratory of Precision and Intelligent Chemistry University of Science and Technology of China Hefei China State Key Laboratory of Magnetic Resonance and Atomic and Molecular Physics Innovation Academy for Precision Measurement Science and Technology Chinese Academy of Sciences Wuhan China Beijing Synchrotron Radiation Facility Institute of High Energy Physics University of Chinese Academy of Sciences Beijing China Key Lab of Biomass Chemical Engineering of Ministry of Education College of Chemical and Biological Engineering Zhejiang University Hangzhou China Interdisciplinary Institute of NMR and Molecular Sciences School of Chemistry and Chemical Engineering Wuhan University of Science and Technology Wuhan China Changping Laboratory Beijing China

Deconstructive transformation of carbon-carbon double bonds (C=C) is a pivotal strategy in synthetic chemistry and drug discovery. Despite the substantial advances in olefin metathesis and ozonolysis for natural product synthesis through C=C double-bond cleavage, the catalytic remodeling of complex molecules through C=C double-bond deconstruction has been underdeveloped. We report a heterogeneous copper-catalyzed C=C double-bond cleavage, which enables the remodeling of complex molecules by converting the carbons on either side of the C=C double bond to carbonyl and cyano groups, respectively. In particular, this method provides an efficient protocol to conveniently transform terpenoids, glycals, steroids, and bioactive molecules to privileged scaffolds with underexplored chemical space.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Exploring Video-Based Driver Activity Recognition under Noisy Labels

arXiv

引用

arXiv 2025年

作者： Fan, Linjuan Wen, Di Peng, Kunyu Yang, Kailun Zhang, Jiaming Liu, Ruiping Chen, Yufan Zheng, Junwei Wu, Jiamin Han, Xudong Stiefelhagen, Rainer Institute for Anthropomatics and Robotics Karlsruhe Institute of Technology Germany School of Robotics National Engineering Research Center of Robot Visual Perception and Control Technology Hunan University Changsha China Shanghai AI Lab China University of Sussex United Kingdom

As an open research topic in the field of deep learning, learning with noisy labels has attracted much attention and grown rapidly over the past ten years. Learning with label noise is crucial for driver distraction behavior recognition, as real-world video data often contains mislabeled samples, impacting model reliability and performance. However, label noise learning is barely explored in the driver activity recognition field. In this paper, we propose the first label noise learning approach for the driver activity recognition task. Based on the cluster assumption, we initially enable the model to learn clustering-friendly low-dimensional representations from given videos and assign the resultant embeddings into clusters. We subsequently perform co-refinement within each cluster to smooth the classifier outputs. Furthermore, we propose a flexible sample selection strategy that combines two selection criteria without relying on any hyperparameters to filter clean samples from the training dataset. We also incorporate a self-adaptive parameter into the sample selection process to enforce balancing across classes. A comprehensive variety of experiments on the public Drive&Act dataset for all granularity levels demonstrates the superior performance of our method in comparison with other label-denoising methods derived from the image classification field. The source code is available at https://***/ilonafan/DAR-noisy-labels. Copyright © 2025, The Authors. All rights reserved.

关键词： Object 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：