检索结果-内蒙古大学图书馆

17th Asian Conference on Computer Vision, ACCV 2024

作者： Huang, Xianliang Chen, Shuhang Zhong, Zhizhou Gou, Jiajie Guan, Jihong Zhou, Shuigeng Shanghai Key Lab of Intelligent Information Processing and School of Computer Science Fudan University Shanghai China School of Computer Science and Technology Tongji University Shanghai China

ISBN: (纸本)9789819609710

Recent developments in Neural Radiance Fields (NeRF) have showcased notable progress in the synthesis of novel views. Nevertheless, there is limited research on inpainting 3D scenes using implicit representations. Traditional approaches utilizing 3D networks for direct 3D inpainting often falter in high-resolution settings, mainly due to GPU memory constraints. This paper introduces Hi-NeRF, an innovative 3D inpainting approach designed to remove arbitrary 3D objects by hybridizing 2D inpainting strategies with NeRF techniques. Recognizing that prevailing 2D inpainting methods often fail to grasp the 3D geometric intricacies of scenes, we leverage the unique capability of NeRF in capturing these structures. Additionally, we propose a multi-view perceptual loss (MVPL) to harness multi-view data, ensuring that 2D inpainting and implicit 3D representations can mutually compensate for each other. Furthermore, we refine the output from the Segment Anything Model (SAM) using image dilation to produce accurate multi-view masks. To finalize the process, we employ Instant-NGP to efficiently retrieve 3D-consistent scenes from 3D-consistent inpainted images. As there is no multi-view 3D scene datasets with corresponding masks, we construct both real-world and synthetic scenes for the multi-view 3D scene inpainting task, which serves as a benchmark dataset. Experimental results on both indoor and outdoor scenes highlight the superiority of our approach over the existing 2D inpainting methods and NeRF-based baselines. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.

关键词： 3D modeling

来源：评论

学校读者我要写书评

暂无评论

Motion Matters: Difference-based Multi-scale Learning for Infrared UAV Detection

Motion Matters: Difference-based Multi-scale Learning for In...

引用

2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops, CVPRW 2023

作者： He, Ruian Zhou, Shili Cheng, Ri Sun, Yuqi Tan, Weimin Yan, Bo Shanghai Collaborative Innovation Center of Intelligent Visual Computing Fudan University School of Computer Science Shanghai Key Laboratory of Intelligent Information Processing Shanghai China

ISBN: (纸本)9798350302493

Unmanned Aerial Vehicle (UAV) detection in the wild is a challenging task due to the presence of background noise and the varying size of the object. To address these obstacles, we propose a novel learning framework for robust UAV detectors, which we call Difference-based Multi-scale Learning (DML). We argue that motion information matters in UAV detection because of the low recognition in one frame. Our method utilizes the frame difference of multiple previous frames, extracting motion information and blocking background noise. We also fuse multiple spatial-temporal scales for training and inferencing, enabling fusion from different sources. In addition, to better evaluate the performance of UAV detection in different scales, we propose Multi-Scale Average Precision (MSAP) metric to aggregate the detection accuracy over multiple scales. Through extensive experiments, we demonstrate that our proposed approach improves the detection accuracy of baseline models. Notably, we achieve SOTA performance in the 3rd Anti-UAV Challenge, with 2nd place in Track 2 and 4th place in Track 1. © 2023 IEEE.

关键词： Antennas

来源：评论

学校读者我要写书评

暂无评论

Uncer2Natural: Uncertainty-Aware Unsupervised Image Denoising 48

Uncer2Natural: Uncertainty-Aware Unsupervised Image Denoisin...

引用

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

作者： Huang, Chenyu Tan, Weimin Shi, Jiaxing Xing, Zhen Yan, Bo Shanghai Collaborative Innovation Center of Intelligent Visual Computing Fudan University School of Computer Science Shanghai Key Laboratory of Intelligent Information Processing Shanghai China

ISBN: (纸本)9781728163277

Recently, unsupervised image denoising methods learning from paired noisy samples have received increasing attention. These methods build on the idea that the mean of multiple noisy images of the same scene is the ideal clean image. However, these methods ignore the effect of Aleatoric uncertainty in the noisy image (e.g., pixels deviating from the expected distribution). The presence of Aleatoric uncertainty causes degradation of the reconstructed target pixels, resulting in high uncertainty for these pixels (i.e., low confidence), which in turn leads to sub-optimal denoising results. To address this problem, we propose a novel uncertainty-aware unsupervised image denoising method named Uncer2Natural (U2N). It dynamically predicts the Aleatoric uncertainty for each noisy sample and produces satisfactory denoising results by reducing the effect of Aleatoric uncertainty. Extensive experimental results show that U2N outperforms state-of-the- art unsupervised image denoising methods in terms of both quantitative metrics and qualitative visual quality. © 2023 IEEE.

关键词： Pixels

来源：评论

学校读者我要写书评

暂无评论

Core-shell nanoparticle enhanced Raman spectroscopy in situ probing the composition and evolution of interfacial species on PtCo surfaces

引用

Nano Research 2024年第6期17卷 4687-4692页

作者： Jing Liu Han-Liang Zhong Xiangyu Li Mu-Fei Yue Wei-Min Yang Xueqiu You Jing-Hua Tian Yao-Hui Wang Jian-Feng Li Fujian Provincial Key Laboratory of Oceanic Information Perception and Intelligent Processing School of Ocean Information EngineeringJimei UniversityXiamen 361021China College of Chemistry Fuzhou UniversityFuzhou 350108China State Key Laboratory of Physical Chemistry of Solid Surfaces MOE Key Laboratory of Spectrochemical Analysis and InstrumentationiChEMCollege of Chemistry and Chemical EngineeringCollege of MaterialsCollege of EnergyXiamen UniversityXiamen 361005China Innovation Laboratory for Sciences and Technologies of Energy Materials of Fujian Province(IKKEM) Xiamen 361005China Shenzhen Research Institute of Xiamen University Shenzhen 518000China

The composition and evolution of interfacial species play a key role during electrocatalytic *** the structural evolution and intermediate during catalytic process by in situ characterization can shed new light on the electrocatalytic reaction mechanism and develop highly efficient ***,directly probing the interfacial species is extremely difficult for most spectroscopic techniques due to complicated interfacial environment and ultra-low surface ***,electrochemical core-shell nanoparticle enhanced Raman spectroscopy is utilized to probe the composition and evolution processes of interfacial species on Au@Pt,Au@Co,and Au@PtCo core-shell nanoparticle *** spectral evidences of interfacial intermediates including hydroxide radical(OH*),superoxide ion(O_(2)^(−)),as well as metal oxide species are directly captured by in situ Raman spectroscopy,which are further confirmed by the both isotopic experiment and density functional theory *** results provide a mechanistic guideline for the rational design of highly efficient electrocatalysts.

关键词： in situ Raman core-shell nanoparticle PtCo interfacial species composition and evolution

来源：评论

学校读者我要写书评

暂无评论

Vortex clusters and their active control in a cold Rydberg atomic system with PT-symmetric Bessel potential

引用

Chinese Physics B 2024年第12期33卷 180-186页

作者： Zhuo Fan Yi Shi Hang Wang Yuan Zhao Wei Peng Siliu Xu Key Laboratory of Optoelectronic Sensing and Intelligent Control Hubei University of Science and TechnologyXianning 437100China School of Biomedical Engineering and Imaging Xianning Medical CollegeHubei University of Science and TechnologyXianning 437100China School of Electronic and Information Engineering Hubei University of Science and TechnologyXianning 437100China

We propose an approach for generating robust two-dimensional(2D)vortex clusters(VCs)in a Rydberg atomic system by utilizing parity-time(PT)symmetric optical Bessel *** show that the system supports novel multicore VCs with four and eight cores,corresponding to topological charges 2 and 4,*** stability of these VCs can be dynamically adjusted through the manipulation of the gain-loss component,Kerr nonlinearities,and the degree of nonlocality inherent in the Rydberg *** VCs are confined within the first lattice well of the Bessel potential,and both the power and width of lights undergo a quasi-periodic breathing phenomenon,which is attributed to the power exchange between the light fields and Bessel *** self-attractive and self-repulsive Kerr interactions can sustain robust VCs within this *** insights presented here not only facilitate the creation and manipulation of 2D VCs through PT-symmetric potentials but also pave the way for potential applications in optical information processing and transmission.

关键词： vortex clusters PT symmetry Bessel potential Rydberg-Rydberg interaction

来源：评论

学校读者我要写书评

暂无评论

Image segmentation with texture-oriented superpixel 2

Image segmentation with texture-oriented superpixel

引用

2nd International Symposium on Computer Applications and information Systems, ISCAIS 2023

作者： Song, Dongke Huang, Heming Zhang, Huiyun School of computer science and technology Qinghai Normal University Xining810008 China State Key Laboratory of Tibetan intelligent information processing and application Xining810008 China

ISBN: (纸本)9781510666665

Superpixel segmentation is an image preprocessing technique that uses pixel blocks instead of pixels to improve the efficiency of subsequent image tasks. Existing methods are not sensitive to image texture. To solve this problem, a texture-oriented superpixel (TOS) segmentation method is proposed. Firstly, an adaptive parameter function based on pixel boundary probability is used to calculate the distance. Secondly, a gradual merging strategy is used for merging. And finally, a loop iterative framework is used to optimize the seeds. The experimental results show that TOS can effectively preserve the smaller texture information, generate seeds that are more consistent with image texture, and better detect image texture. © 2023 SPIE.

关键词： Superpixels

来源：评论

学校读者我要写书评

暂无评论

NCCNet: Arbitrary Neural Style Transfer with Multi-channel Conversion 1

引用

12th International Conference on Image and Graphics, ICIG 2023

作者： Xiang, Jin Zhao, Huihuang Lin, Mugang Liu, Qingyun College of Computer Science and Technology Hengyang Normal University Hengyang421002 China Hunan Provincial Key Laboratory of Intelligent Information Processing and Application Hengyang421002 China

ISBN: (数字)9783031463143

ISBN: (纸本)9783031463136

Arbitrary style transfer aims to create a novel image from a content image and a style image, stylizing the content image with the style of the style image. However, recent algorithms are prone to unnatural output due to focusing on the statistical distribution of deep features and ignoring the relationship between channels in feature maps. In this paper, we devise a novel Arbitrary Neural Style Transfer Network called NCCNet, which is useful for artistic style transfer. Specifically, NCCNet creates good visual effects by fusing different layers of textures. Moreover, we design an Effective Channel Conversion Modules (ECCM) which uses a multi-channel matrix to rearrange style feature maps to the appropriate range by analyzing the relationship between channels. Then, we can obtain a covariance based on the rearranged style feature map. At the same time, we design a decoder named Self-Adaptive Decoder (SA-Decoder) that conforms to our algorithm. Moreover, we apply our algorithm to the video style transfer and get good results. Qualitative and quantitative evaluations demonstrate that NCCNet performs well in arbitrary video and image style transfer tasks. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2023.

关键词： Textures

来源：评论

学校读者我要写书评

暂无评论

Tackling Sparse Facts for Temporal Knowledge Graph Completion 25

Tackling Sparse Facts for Temporal Knowledge Graph Completio...

引用

34th ACM Web Conference, WWW 2025

作者： Zhang, Yuchao Kong, Xiangjie Ye, Kailun Shen, Guojiang Zheng, Shangfei Zhejiang University of Technology College of Computer Science and Technology Hangzhou China Zhejiang Key Laboratory of Visual Information Intelligent Processing Hangzhou China Zhejiang Sci-Tech University School of Computer Science and Technology Hangzhou China

ISBN: (纸本)9798400712746

Temporal knowledge graph completion (TKGC) seeks to develop more comprehensive knowledge representations by addressing missing relationships and entities within temporal knowledge graphs (TKGs), thereby enhancing reasoning and predictive capabilities in downstream tasks. Nonetheless, real-world knowledge—such as the progression of social network interactions and the unfolding of news events—is inherently dynamic, resulting in substantial sparsity issues in TKGs that profoundly impair the performance of TKGC models. To overcome this challenge, we introduce the Adaptive Neighborhood Enhancement Layer (ANEL), a novel module that can be effortlessly integrated into existing TKGC models to substantially elevate the representation quality of sparse entities. ANEL first derives initial entity embeddings through a base model and then uncovers concealed semantic relationships between entities via a latent relation module, enriching the explicit relationships within the knowledge graph. Furthermore, ANEL incorporates an adaptive latent information adjustment component, which dynamically calibrates the influence of latent information based on the entity’s relational structure: entities with fewer connections derive greater benefit from latent information, while entities with denser connections become less dependent on latent augmentation, ensuring precise and resilient representations. We conducted comprehensive experiments on four prominent benchmark datasets, and the results underscore the effectiveness and superiority of ANEL in TKGC tasks. © 2025 Copyright held by the owner/author(s).

关键词： Knowledge graph

来源：评论

学校读者我要写书评

暂无评论

A semantic segmentation method for UAV image based on improved BiseNetv2 5

A semantic segmentation method for UAV image based on improv...

引用

5th International Conference on Artificial Intelligence and Computer Science, AICS 2023

作者： Guo, Fengjuan Han, Sen Zhao, Jiahang Han, Jiale Xin, Jing Shaanxi Key Laboratory of Integrated and Intelligent Navigation The 20th Research Institute of China Electronics Technology Corporation Shaanxi Xi’an710068 China Shaanxi Key Laboratory of Complex System Control and Intelligent Information Processing Xi’an University of Technology Shaanxi Xi’an710048 China

ISBN: (纸本)9781510668621

Aiming at the problem of low semantic segmentation accuracy of tiny targets in UAV aerial images, a bilateral semantic segmentation network T-BiseNetv2 is proposed. The proposed segmentation network is based on BiseNetv2, and the Transformer structure is introduced into semantic branch, and the detail branch and feature fusion module are reconstructed, which improves the network’s capacity to capture both global context information and local semantic information. Several semantic segmentation experiments are conducted on the standard semantic segmentation dataset UAVID, experimental results show that the proposed semantic segmentation network T-BiseNetv2 has a MIoU of 69.52%, which is about 2.2% higher than the original BiseNetv2 network. Especially, for "Human", the IoU of TBiseNetv2 is about 4.3% higher than that of the popular UNetFormer. © 2023 SPIE.

关键词： Unmanned aerial vehicles (UAV)

来源：评论

学校读者我要写书评

暂无评论

Multiagent trajectory prediction with global-local scene-enhanced social interaction graph network

Multiagent trajectory prediction with global-local scene-enh...

引用

作者： Lin, Xuanqi Zhang, Yong Wang, Shun Piao, Xinglin Yin, Baocai Beijing Key Laboratory of Multimedia and Intelligent Software Technology Beijing Artificial Intelligence Institute Faculty of Information Technology Beijing University of Technology Beijing China

Trajectory prediction is essential for intelligent autonomous systems like autonomous driving, behavior analysis, and service robotics. Deep learning has emerged as the predominant technique due to its superior modeling capability for trajectory data. However, deep learning-based models face challenges in effectively utilizing scene information and accurately modeling agent interactions, largely due to the complexity and uncertainty of real-world scenarios. To mitigate these challenges, this study presents a novel multiagent trajectory prediction model, termed the global-local scene-enhanced social interaction graph network (GLSESIGN), which incorporates two pivotal strategies: global-local scene information utilization and a social adaptive attention graph network. The model hierarchically learns scene information relevant to multiple intelligent agents, thereby enhancing the understanding of complex scenes. Additionally, it adaptively captures social interactions, improving adaptability to diverse interaction patterns through sparse graph structures. This model not only improves the understanding of complex scenes but also accurately predicts future trajectories of multiple intelligent agents by flexibly modeling intricate interactions. Experimental validation on public datasets substantiates the efficacy of the proposed model. This research offers a novel model to address the complexity and uncertainty in multiagent trajectory prediction, providing more accurate predictive support in practical application scenarios. © 2024 John Wiley & Sons Ltd.

关键词： Multi agent systems

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：