检索结果-内蒙古大学图书馆

32nd ACM International Conference on Multimedia, MM 2024

作者： Zhang, Zhenyu Chen, Guangyao Zou, Yixiong Li, Yuhua Li, Ruixuan School of Computer Science and Technology Huazhong University of Science and Technology Wuhan China National Key Laboratory for Multimedia Information Processing School of Computer Science Peking University Beijing China

ISBN: (纸本)9798400706868

Few-shot open-set recognition (FSOR) is a challenging task that requires a model to recognize known classes and identify unknown classes with limited labeled data. Existing approaches, particularly Negative-Prototype-Based methods, generate negative prototypes based solely on known class data. However, as the unknown space is infinite while the known space is limited, these methods suffer from limited representation capability. To address this limitation, we propose a novel approach, termed Diversified Negative Prototypes Generator (DNPG), which adopts the principle of "learning unknowns from unknowns."Our method leverages the unknown space information learned from base classes to generate more representative negative prototypes for novel classes. During the pre-training phase, we learn the unknown space representation of the base classes. This representation, along with inter-class relationships, is then utilized in the meta-learning process to construct negative prototypes for novel classes. To prevent prototype collapse and ensure adaptability to varying data compositions, we introduce the Swap Alignment (SA) module. Our DNPG model, by learning from the unknown space, generates negative prototypes that cover a broader unknown space, thereby achieving state-of-the-art performance on three standard FSOR datasets. The repository of this project is available at https://***/iCGY96/DNPG. © 2024 ACM.

关键词： Zero-shot learning

来源：评论

学校读者我要写书评

暂无评论

DIR-YOLOv5: A real-time Drone-perspective Infrared object detection method based on YOLOv5

DIR-YOLOv5: A real-time Drone-perspective Infrared object de...

引用

2023 Applied Optics and Photonics China: Infrared Devices and Infrared technology;and Terahertz technology and Applications, AOPC 2023

作者： Yuexing, Wang Chuangang, Xu Jianzhong, Su Jinwen, Tian National Key Laboratory of Multispectral Information Intelligent Processing Technology School of Artificial Intelligence and Automation Huazhong University of Science and Technology Wuhan430074 China Tianjin Jinhang Institute of Technical Physics Tianjin300308 China

ISBN: (纸本)9781510672260

With the advancement of drone technology, object detection from the perspective of drones has found extensive applications in various fields, including surveillance, search operations, and reconnaissance tasks. Currently, most drones in the market are equipped with visible light imagers, while some high-end drones are equipped with infrared imaging detectors capable of performing infrared object detection tasks. Infrared imaging utilizes a passive imaging mode, enabling it to detect thermal radiation emitted by objects. As a result, it offers the distinct advantage of continuous operation without being restricted by daylight conditions. In comparison to visible imaging, infrared imaging uses longer wavelengths and possesses a certain level of penetration capability through clouds and smoke. Consequently, infrared object detection represents a significant research area within the field of object detection. However, detecting infrared objects, especially small ones, remains challenging due to the complexity of background information, lower resolution compared to visible images, and the lack of shape and texture information in infrared images. In response to these challenges, this study proposes a real-time drone-perspective infrared (IR) object detection method based on the YOLOv5 framework, known as DIR-YOLOv5. To effectively address the challenge of infrared vehicles occupying fewer pixels in the drone's perspective image and making objects difficult to detect, the coordinate attention (CA) for feature enhancement is introduced. we also introduce a Spatial-Channel dynamic and query-aware sparse attention mechanism (SCBiFormer), which is optimized based on BiFormer. Additionally, we redefine the loss function as the Repulsion Loss function to tackle the problem of infrared vehicle objects gathering and overlapping occlusion in scenarios like parking lots. Furthermore, we expand the ISVD infrared image object detection dataset to include multiple scenarios and conduct

关键词： Image enhancement

来源：评论

学校读者我要写书评

暂无评论

Infrared Dim Small Target Detection Algorithm Based on Adaptive Attention Transformer Network

引用

Journal of Shanghai Jiaotong University (science) 2025年 1-13页

作者： Ding, Leqi Xia, Yilin Wang, Biyun Cai, Yunze Department of Automation Shanghai Jiao Tong UniversityKey Laboratory of System Control and Information Processing of Ministry of Education Shanghai200240 China National Key Laboratory of Air-based Information Perception and Fusion Henan Luoyang471009 China Shanghai Institute of Aerospace Control Technology Infrared Detection Technology R & D Center of China Aerospace Science and Technology Corporation Shanghai201109 China

This paper presents a novel algorithm for detecting infrared dim small targets in complex backgrounds by integrating both global and local feature information. The proposed method employs an adaptive attention transformer network which addresses challenges associated with computational complexity and information redundancy in traditional self-attention mechanisms. A global context aggregation module is designed to address the loss of small target features in high-level feature maps. This module compensates for spatial distribution information during the multi-scale feature fusion process, ensuring the preservation of target features. Comparison experiments were conducted on two datasets, NUAA-SIRST and IRSTD-1k. The proposed algorithm achieved 0.784 for mIoU and 0.878 for F-measure on NUAA-SIRST, 0.632 for mIoU and 0.807 for F-measure on IRST-1k, all of which are the highest values. The proposed algorithm shows a significant improvement over other existing detection algorithms, demonstrating its capability to accurately extract both local and global features of targets with high recall rates and low false alarm rates across various scenarios. The ablation study results highlight the significant improvements contributed by the key modules to the proposed algorithm. (Figure presented.) © Shanghai Jiao Tong University 2025.

关键词： A adaptive attention global context aggregation module infrared dim small target detection TP391.4 Transformer network

来源：评论

学校读者我要写书评

暂无评论

Differentially Private Federated Stochastic Primal-Dual Learning for Internet of Vehicles

引用

IEEE Internet of Things Journal 2025年第11期12卷 17034-17050页

作者： Li, Yiwei Wang, Shuai Chang, Tsung-Hui Xiamen University of Technology Fujian Key Laboratory of Communication Network and Information Processing Xiamen361024 China University of Electronic Science and Technology of China National Key Laboratory of Wireless Communications Chengdu611731 China Chinese University of Hon g Kong School of Science and Engineering Shenzhen518172 China Shenzhen Research Institute of Big Data Shenzhen518172 China

Federated learning (FL) has the potential to empower Internet of Vehicles (IoV) networks by enabling smart vehicles (SVs) to participate in the learning process under the orchestration of a vehicular service provider while keeping data locally. In this article, we propose a novel federated stochastic primal-dual algorithm with differential privacy (FedSPD-DP) to ensure robust privacy protection for FL based IoV (FL-IoV) systems. The FedSPD-DP algorithm leverages multiple steps of local stochastic gradient descent (SGD) and partial client participation (PCP) to improve communication efficiency while incorporating differential privacy (DP) to ensure privacy protection. Our theoretical analysis explores the impact of these strategies on learning performance. Specifically, we demonstrate that the data sampling strategy and PCP enhance data privacy, while a larger number of local SGD steps may increase privacy leakage, revealing a nontrivial tradeoff between communication efficiency and privacy protection. Extensive experiments on real-world data validate the effectiveness of the proposed algorithm, showing superior performance compared to state-of-the-art FL algorithms, and confirming the analytical results and properties. © 2014 IEEE.

关键词： Federated learning

来源：评论

学校读者我要写书评

暂无评论

Object-Aware Video Matting with Cross-Frame Guidance

arXiv

引用

arXiv 2025年

作者： Zhang, Huayu Wu, Dongyue Shao, Yuanjie Sang, Nong Gao, Changxin National Key Laboratory of Multispectral Information Intelligent Processing Technology School of Artificial Intelligence and Automation Huazhong University of Science and Technology China

Recently, trimap-free methods have drawn increasing attention in human video matting due to their promising performance. Nevertheless, these methods still suffer from the lack of deterministic foreground-background cues, which impairs their ability to consistently identify and locate foreground targets over time and mine fine-grained details. In this paper, we present a trimap-free Object-Aware Video Matting (OAVM) framework, which can perceive different objects, enabling joint recognition of foreground objects and refinement of edge details. Specifically, we propose an Object-Guided Correction and Refinement (OGCR) module, which employs cross-frame guidance to aggregate object-level instance information into pixel-level detail features, thereby promoting their synergy. Furthermore, we design a Sequential Foreground Merging augmentation strategy to diversify sequential scenarios and enhance capacity of the network for object discrimination. Extensive experiments on recent widely used synthetic and real-world benchmarks demonstrate the state-of-the-art performance of our OAVM with only an initial coarse mask. The code and model will be available. © 2025, CC BY.

关键词： Object recognition

来源：评论

学校读者我要写书评

暂无评论

Cross-Domain Detection Transformer with Multi-view Adaptive Feature Alignment in Remote Sensing Imagery 2

Cross-Domain Detection Transformer with Multi-view Adaptive ...

引用

2nd IEEE International Conference on Signal, information and Data processing, ICSIDP 2024

作者： Wang, Shu Han, Jianhong Wang, Ying Hao, Xinyuan Luo, Zhaoyi Wang, Yupei Chen, Liang Beijing Institute of Technology National Key Laboratory for Science and Technology on Space-Born Intelligent Information Processing Beijing China Shanghai Academy of Spaceflight Technology Shanghai China

ISBN: (纸本)9798331515669

Unsupervised Domain Adaptation (UDA) techniques are crucial for remote sensing object detection, designed to address performance degradation caused by the domain gap between training and test data. These methods leverage unlabeled target domain data, thus alleviating the high costs associated with data annotation. Recent developments in Detection Transformers (DETR) have simplified the detection pipeline and attracted significant research interest. Building on this architecture, we introduce an unsupervised domain adaptation detector for remote sensing object detection. Specifically, we introduce a multi-view adaptive feature alignment module that initially captures domain-specific features in complex backgrounds by leveraging a cross-attention mechanism. Subsequently, we employ contrastive learning to enforce the aggregation of domain-specific features from various perspectives, thereby improving the accuracy of feature alignment. Moreover, we demonstrate that integrating the self-training framework into DETR-based detectors can significantly mitigate the domain gap by further utilizing unlabeled data in the target domain. We validated the effectiveness and generalizability of our method across two remote sensing cross-domain detection scenarios using four public datasets. © 2024 IEEE.

关键词： object detection remote sensing imagery Unsupervised domain adaptation

来源：评论

学校读者我要写书评

暂无评论

LED: A Large-scale Real-world Paired Dataset for Event Camera Denoising

LED: A Large-scale Real-world Paired Dataset for Event Camer...

引用

Conference on Computer Vision and Pattern Recognition (CVPR)

作者： Yuxing Duan National Key Lab of Multispectral Information Intelligent Processing Technology Huazhong University of Science and Technology

ISBN: (数字)9798350353006

ISBN: (纸本)9798350353013

Event camera has significant advantages in capturing dynamic scene information while being prone to noise in-terference, particularly in challenging conditions like low threshold and low illumination. However, most existing research focuses on gentle situations, hindering event camera applications in realistic complex scenarios. To tackle this limitation and advance the field, we construct a new paired real-world event denoising dataset (LED), including 3K sequences with 18K seconds of high-resolution (1200*680) event streams and showing three notable distinctions compared to others: diverse noise levels and scenes, larger-scale with high-resolution, and high-quality GT. Specifically, it contains stepped parameters and varying illumination with diverse scenarios. Moreover, based on the property of noise events inconsistency and signal events consistency, we propose a novel effective denoising frame-work(DED) using homogeneous dual events to generate the GT with better separating noise from the raw. Further-more, we design a bio-inspired baseline leveraging Leaky-Integrate-and-Fire (LIF) neurons with dynamic thresholds to realize accurate denoising. The experimental results demonstrate that the remarkable performance of the proposed approach on different datasets. The dataset and code are at https://***/Yee-Sing/led.

关键词： Computer vision Noise reduction Noise Neurons Lighting Light emitting diodes Cameras

来源：评论

学校读者我要写书评

暂无评论

Anisotropic optical path model for characterizing statistics of scattered field

引用

Optics Express 2025年第10期33卷 21306-21317页

作者： Bao, Ke Li, Yuegang Nie, Zhenwu Sun, Shuai Liu, Weitao Institute for Quantum Science and Technology College of Science National University of Defense Technology Changsha410073 China Interdisciplinary Center for Quantum Information National University of Defense Technology Changsha410073 China Hunan Key Laboratory of Mechanism and Technology of Quantum Information Changsha410073 China State Key Laboratory of Photonics and Communications Institute for Quantum Sensing and Information Processing Shanghai Jiao Tong University Shanghai200240 China Hunan Research Center of the Basic Discipline for Physical States National University of Defense Technology Changsha410073 China

A model that captures the statistical properties of the scattered light field can offer a physical explanation and computational support for designing and evaluating imaging schemes. Here, we propose an Anisotropic Optical Path (AOP) model that incorporates statistics of different optical paths affected by scattering, with photons involved in different processes. The model enables simulation of a limited field of view (FOV) constrained by the memory effect. It also reflects spectral differences under broadband illumination, thereby enhancing its applicability for light field simulation in higher degrees of freedom. Due to the decorrelation trend of speckle patterns at different wavelengths, the model verifies the feasibility of using scattering media as spectral filters. © 2025 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement.

关键词： Speckle

来源：评论

学校读者我要写书评

暂无评论

Observer-based Hierarchical Control of Heterogeneous Multiagent Systems Under Attacks and Disturbances

Observer-based Hierarchical Control of Heterogeneous Multiag...

引用

第43届中国控制会议

作者： Yongran Zhi Huijin Fan Lei Liu Bo Wang The National Key Laboratory of Multispectral Information Intelligent Processing Technology School of Artificial Intelligence and Automation Huazhong University of Science and Technology

ISBN: (数字)9789887581581

ISBN: (纸本)9798350366907

The resilient cooperative output regulation of heterogeneous leader-follower multiagent systems(MASs) under asynchronous DoS attacks and unknown external disturbances is investigated in this paper. An observer-based hierarchical control frame is constructed, which contains communication layer and controller layer. In the communication layer, based on an improved dynamic event-triggered mechanism, a novel fully distributed observer is proposed to estimate the leader's state under the asynchronous DoS attack. In the controller layer, an anti-disturbance state observer is firstly designed to estimate the follower's state in the existence of unknown external disturbances, then an output feedback controller is constructed to track the estimate leader's state. The proposed two observers has increased the tolerance to DoS attacks and external disturbances. Finally,simulation results have demonstrated the effectiveness of the proposed hierarchical control frame.

关键词： Fully distributed event-triggered observer hierarchical control asynchronous Dos attacks unknown external disturbances heterogeneous MASs

来源：评论

学校读者我要写书评

暂无评论

RIS-Aided Cell-Free Massive MIMO Systems With Low-Resolution ADCs: Uplink Performance Analysis and Optimization

引用

IEEE Internet of Things Journal 2025年第11期12卷 18249-18265页

作者： Xiong, Youzhi Tang, Liu Sun, Sanshan Yang, Songjie Liu, Li Mao, Sun Zhang, Zhongpei Wei, Ning Sichuan Normal University Laboratory of Information Intelligent Processing and Communications College of Physics and Electronic Engineering Chengdu610101 China University of Electronic Science and Technology of China National Key Laboratory of Wireless Communications Chengdu611731 China Sichuan Normal University College of Computer Science Chengdu610101 China

This article investigates the uplink performance of reconfigurable intelligent surface (RIS)-aided cell-free (CF) massive multiple-input-multiple-output (mMIMO) systems over spatially correlated Rayleigh fading channels. We consider multiple RISs and low-resolution analog-to-digital converters (ADCs) to improve the system energy efficiency (EE). We first provide an aggregated channel estimation technique with less pilot overhead. By exploiting the statistical channel state information (CSI), we further optimize the RISs’ phase shifts with the goal of minimizing the total normalized mean square error (NMSE) of the estimated aggregated channels. Subsequently, we derive the closed-form expression of the uplink spectral efficiency (SE) for quantization-aware minimum mean-square error (MMSE) combining. Third, based on the closed-form SE expression and power consumption model, we formulate and solve an optimization problem that maximizes the uplink EE under the constraints of transmit power and total ADC quantization bits. Specifically, by leveraging the Dinkelbach transform, Lagrangian dual transform, and fractional programming (FP) techniques, an alternating optimization (AO)-based algorithm is proposed to jointly obtain the bit allocation (BA) scheme among all access points (APs) and the uplink power control (PC) strategy for all users. Finally, numerical results validate the correctness of the closed-form SE expression and show the effectiveness of the proposed optimization methods for phase shift design and EE maximization. © 2014 IEEE.

关键词： Mean square error

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：