检索结果-内蒙古大学图书馆

A multi-level multi-attention mechanism millimeter-wave radar and camera fusion method for 3d object detection

SIGNAL IMAGE ANd VIdEO PROCESSING 2025年第6期19卷 1-11页

作者： Miao, Zehua Li, Yinbei Wu, Zizhuo Yang, Jiaqiang Ma, Yuliang Hangzhou Dianzi Univ Sch Automat Hangzhou 310018 Peoples R China Peoples Hosp Huaiyin Jinan Key Lab Rehabil & Evaluat Motor Dysfunct Jinan 250100 Shandong Peoples R China Zhejiang Univ Coll Elect Engn Hangzhou 310027 Peoples R China Shandong BetR Med Technol Co Ltd Jinan 250100 Shandong Peoples R China

In the field of autonomous driving, a commonly employed method to enhance detection accuracy and robustness is the fusion of multi-sensor perception. The fusion of millimeter-wave radar and camera can effectively complement each other, providing sufficient semantic information while ensuring robustness against varying illumination and weather conditions, at a lower cost. In this paper, we focus on the fusion of millimeter-wave radar point cloud features and image features, proposing a multi-level multi-attention feature-level fusion method. By improving the dLA34 backbone network to expand the model's receptive field, we fuse point cloud features at multiple levels with image features and utilize an improved feature pyramid to handle features of both modalities, ensuring good cross-channel information capture capability. Our model leverages the advantages of multi-level multi-attention, achieving an accuracy of 34.3% in the challenging nuScenes dataset, which is 1.9% higher than the second-best model, demonstrating promising performance.

关键词： Attention mechanism Multi-level feature fusion 3d object detection Radar

来源：评论

学校读者我要写书评

暂无评论

Multi-Modal Fusion Based on depth Adaptive Mechanism for 3d object detection

引用

IEEE TRANSACTIONS ON MULTIMEdIA 2025年 27卷 707-717页

作者： Liu, Zhanwen Cheng, Juanru Fan, Jin Lin, Shan Wang, Yang Zhao, Xiangmo Changan Univ Sch Informat Engn Xian 710064 Peoples R China Univ Sci & Technol China Sch Informat Hefei 230026 Peoples R China

Lidars and cameras are critical sensors for 3d object detection in autonomous driving. despite the increasing popularity of sensor fusion in this field, accurate and robust fusion methods are still under exploration due to non-homogenous representations. In this paper, we find that the complementary roles of point clouds and images vary with depth. An important reason is that the point cloud appearance changes significantly with increasing distance from the Lidar, while the image's edge, color, and texture information are not sensitive to depth. To address this, we propose a fusion module based on the depth Attention Mechanism (dAM), which mainly consists of two operations: gated feature generation and point cloud division. The former adaptively learns the importance of bimodal features without additional annotations, while the latter divides point clouds to achieve differential fusion of multi-modal features at different depths. This fusion module can enhance the representation ability of original features for different point sets and provide more comprehensive features by using the dual splicing strategy of concatenation and index connection. Additionally, considering point density as a feature and its negative correlation with depth, we build an Adaptive Threshold Generation Network (ATGN) to generate the depth threshold by extracting density information, which can divide point clouds more reasonably. Experiments on the KITTI dataset demonstrate the effectiveness and competitiveness of our proposed models.

关键词： Point cloud compression Three-dimensional displays Feature extraction object detection Laser radar Image color analysis detectors deep learning 3d object detection sensor fusion Lidar sensor camera sensor

来源：评论

学校读者我要写书评

暂无评论

dART3d: depth-Aware Robust Adversarial Training for Monocular 3d object detection

引用

ELECTRONICS LETTERS 2025年第1期61卷

作者： Ju, Xinrui Shang, Xiaoke Li, Xingyuan Ren, Bohua Dongbei Univ Finance & Econ Dalian Peoples R China Dalian Univ Technol Dalian Peoples R China

Monocular 3d object detection plays a pivotal role in the field of autonomous driving and numerous deep learning-based methods have made significant breakthroughs in this area. despite the advancements in detection accuracy and efficiency, these models tend to fail when faced with adversarial attacks, rendering them ineffective. Therefore, bolstering the adversarial robustness of 3d detection models has become a critical issue. To mitigate this issue, we propose a depth-aware robust adversarial training method for monocular 3d object detection, dubbed dART3d. Specifically, we first design an adversarial attack that iteratively degrades the 2d and 3d perception capabilities of 3d object detection models (iterative deterioration of perception), serving as the foundation for our subsequent defense mechanism. In response to this attack, we propose an uncertainty-based residual learning method for adversarial training. Our adversarial training leverages inherent uncertainty to boost robustness against attacks while incorporating depth-aware information enhances resistance to perturbations in both 2d and 3d domains. We conducted extensive experiments on the KITTI 3d dataset, showing that dART3d outperforms direct adversarial training in 3d object detection APR(40) for the car category, with improvements of 4.415%, 4.112% and 3.195% in easy, moderate and hard settings, respectively.

关键词： Adversarial attack 3d object detection monocular vision

来源：评论

学校读者我要写书评

暂无评论

dispersion Adaptive Convolution: Robust Multi-Modal 3d object detection by Incorporating Sensor Characteristics

引用

JOURNAL OF CIRCUITS SYSTEMS ANd COMPUTERS 2025年第7期34卷

作者： Chen, Yaqing Wang, Huaming Nanjing Univ Aeronaut & Astronaut Coll Mech & Elect Engn Nanjing 210016 Jiangsu Peoples R China

Multi-modal three-dimensional (3d) roadside object detection is a challenging yet critical topic for Vehicle-Infrastructure Cooperated Autonomous driving (VICAd). Recently, the Birds-Eye View (BEV) framework has emerged as a promising solution. However, current approaches that process BEV features with isotropic convolutions fail to account for the error characteristics of sensors, leading to inefficiency and instability in disturbed scenarios. This paper addresses this issue by developing the Sensor Characteristics Adaptive Bird's Eye View (SCA-BEV), a simple yet effective roadside 3d object detection framework based on improved deformable convolution and anisotropic attention mechanism. Specifically, we designed the error dispersion Adaptive Convolution (dAC) to enhance the network's anti-interference ability. Moreover, a novel axial attention mechanism was proposed to improve multi-modality fusion. Additionally, we introduce a knowledge distillation framework to boost the performance in camera-only scenarios. The proposed method was validated on the dAIR-V2X dataset and compared against state-of-the-art networks. The experimental results demonstrate that our method achieves higher detection accuracy with remarkable real-time performance in disturbed scenarios, suggesting a broad application prospect in intelligent roadside systems.

关键词： ITS 3d object detection deformable convolutional axial attention

来源：评论

学校读者我要写书评

暂无评论

Cross-domain Generalization for LidAR-Based 3d object detection in Infrastructure and Vehicle Environments

引用

SENSORS 2025年第3期25卷 767-767页

作者： Zhi, Peng Jiang, Longhao Yang, Xiao Wang, Xingzheng Li, Hung-Wei Zhou, Qingguo Li, Kuan-Ching Ivanovic, Mirjana Lanzhou Univ Sch Informat Sci & Engn Lanzhou 730000 Peoples R China Providence Univ Dept Comp Sci & Informat Engn Taichung 43301 Taiwan Univ Novi Sad Fac Sci Novi Sad 21000 Serbia

In the intelligent transportation field, the Internet of Things (IoT) is commonly applied using 3d object detection as a crucial part of Vehicle-to-Everything (V2X) cooperative perception. However, challenges arise from discrepancies in sensor configurations between vehicles and infrastructure, leading to variations in the scale and heterogeneity of point clouds. To address the performance differences caused by the generalization problem of 3d object detection models with heterogeneous LidAR point clouds, we propose the dual-Channel Generalization Neural Network (dCGNN), which incorporates a novel data-level downsampling and calibration module along with a cross-perspective Squeeze-and-Excitation attention mechanism for improved feature fusion. Experimental results using the dAIR-V2X dataset indicate that dCGNN outperforms detectors trained on single datasets, demonstrating significant improvements over selected baseline models.

关键词： 3d object detection V2X cooperative perception LidAR point clouds infrastructure sensors

来源：评论

学校读者我要写书评

暂无评论

SOFW: A Synergistic Optimization Framework for Indoor 3d object detection

引用

IEEE TRANSACTIONS ON MULTIMEdIA 2025年 27卷 637-651页

作者： dai, Kun Jiang, Zhiqiang Xie, Tao Wang, Ke Liu, dedong Fan, Zhendong Li, Ruifeng Zhao, Lijun Omar, Mohamed Harbin Inst Technol State Key Lab Robot & Syst Harbin 150006 Peoples R China State Key Yangtze River Delta HIT Robot Technol Re Wuhu 241000 Peoples R China

In this work, we observe that indoor 3d object detection across varied scene domains encompasses both universal attributes and specific features. Based on this insight, we propose SOFW, a synergistic optimization framework that investigates the feasibility of optimizing 3d object detection tasks concurrently spanning several dataset domains. The core of SOFW is identifying domain-shared parameters to encode universal scene attributes, while employing domain-specific parameters to delve into the particularities of each scene domain. Technically, we introduce a set abstraction alteration strategy (SAAS) that embeds learnable domain-specific features into set abstraction layers, thus empowering the network with a refined comprehension for each scene domain. Besides, we develop an element-wise sharing strategy (ESS) to facilitate fine-grained adaptive discernment between domain-shared and domain-specific parameters for network layers. Benefited from the proposed techniques, SOFW crafts feature representations for each scene domain by learning domain-specific parameters, whilst encoding generic attributes and contextual interdependencies via domain-shared parameters. Built upon the classical detection framework VoteNet without any complicated modules, SOFW delivers impressive performances under multiple benchmarks with much fewer total storage footprint. Additionally, we demonstrate that the proposed ESS is a universal strategy and applying it to a voxels-based approach TR3d can realize cutting-edge detection accuracy on all S3dIS, ScanNet, and SUN RGB-d datasets.

关键词： Three-dimensional displays Feature extraction object detection Point cloud compression Optimization Accuracy Shape Proposals Sun Kernel 3d object detection synergistic optimization deep learning

来源：评论

学校读者我要写书评

暂无评论

MdFusion: Multi-dimension Semantic-Spatial Feature Fusion for LidAR-Camera 3d object detection

引用

REMOTE SENSING 2025年第7期17卷 1240-1240页

作者： Qiao, Renzhong Yuan, Hao Guan, Zhenbo Zhang, Wenbo Xidian Univ Sch Elect Engn Xian 710071 Peoples R China CETC 54th Res Inst Shijiazhuang 050081 Peoples R China Hebei Key Lab Intelligent Informat Percept & Proc Shijiazhuang 050081 Peoples R China

Accurate 3d object detection is becoming increasingly vital for the development of robust perception systems, particularly in applications such as autonomous driving vehicles and robotic systems. Many existing approaches rely on bird's eye view (BEV) feature maps to facilitate multi-modal interaction, as BEV representations enable efficient operations. However, the inherent sparsity of LidAR BEV features often leads to misalignment with the dense semantic information in camera images, resulting in suboptimal fusion quality and degraded detection performance, especially in complex and dynamic environments. To mitigate these issues, this paper proposes a novel multi-dimension semantic-spatial feature fusion (MdFusion) method that combines LidAR and image features in 2d and 3d spaces. Specifically, image semantic features are extracted using the deepLabV3 segmentation network, which captures rich contextual information and is aligned with LidAR point cloud voxel features through a summation operation to achieve precise semantic fusion. Additionally, LidAR BEV features are fused with downsampled image features in 2d space via concatenation and spatially adaptive dilated convolution. The mechanism dynamically adjusts to the spatial characteristics of the data, ensuring robust feature integration. Extensive experiments on the KITTI and ONCE datasets demonstrate that our method achieves competitive performance in complex scenes, significantly improving the multi-modal fusion quality and detection accuracy while maintaining computational efficiency.

关键词： 3d object detection multi-dimension feature fusion multi-modal

来源：评论

学校读者我要写书评

暂无评论

PEPillar: a point-enhanced pillar network for efficient 3d object detection in autonomous driving

引用

VISUAL COMPUTER 2025年第3期41卷 1777-1788页

作者： Sun, Libo Li, Yifan Qin, Wenhu Southeast Univ Sch Instrument Sci & Engn Nanjing 210096 Peoples R China

Pillar-based 3d object detection methods outperform traditional point-based and voxel-based methods in terms of speed. However, most of recent methods in this category use simple aggregation techniques to construct pillar feature maps, which leads to a significant loss of raw point cloud detail and a decrease in detection accuracy. Given the critical demand for both rapid response and high precision in autonomous driving, we introduce PEPillar, an innovative 3d object detection method that adopts point cloud data fusion. Concretely, we firstly use the Point-Enhanced Pillar module to learn pillar and keypoints features from the input data. Then attention mechanism is employed to seamlessly integrate features from multiple sources, which improves the model's ability to detect various objects and demonstrates robustness in complex scenarios. Benefiting from the simplicity of pillar representation, PEPillar can use established 2d convolutional neural networks to solve the challenges in backbone network redesign. The Multi-Receptive Field Neck is introduced to enhance the detection accuracy of smaller objects. Additionally, we design the model into a faster single-stage and a more precise two-stage format to meet various requirements. The results of the evaluation indicate a 5.14% improvement of our method compared to the baseline model in the moderately difficult car detection task, achieving levels comparable to state-of-the-art methods that use point and voxel representations.

关键词： 3d object detection Feature fusion Point clouds Pillar

来源：评论

学校读者我要写书评

暂无评论

STFNET: Sparse Temporal Fusion for 3d object detection in LidAR Point Cloud

引用

IEEE SENSORS JOURNAL 2025年第3期25卷 5866-5877页

作者： Meng, Xin Zhou, Yuan Ma, Jun Jiang, Fangdi Qi, Yongze Wang, Cui Kim, Jonghyuk Wang, Shifeng Changchun Univ Sci & Technol Sch Optoelect Engn Changchun 130022 Peoples R China Leapmotor Hangzhou 310000 Peoples R China Changchun Univ Sci & Technol Zhongshan Inst Zhongshan 528400 Peoples R China Naif Arab Univ Secur Sci Ctr Excellence Cybercrimes & Digital Forens Riyadh 11452 Saudi Arabia

In autonomous driving and robotics, 3d object detection using LidAR point clouds is a critical task. However, existing single-frame 3d object detection methods face challenges such as noise, occlusions, and sparsity, which degrade detection performance. To address these, we propose the sparse temporal fusion network (STFNet), which leverages multiframe historical information to improve 3d object detection accuracy. The contribution of STFNet contains three core modules: multihistory feature alignment module (MFAM), sparse feature extraction module (SFEM), and temporal fusion transformer (TFformer). MFAM: Ego-motion is used for compensation to align frames, establishing correlations between adjacent frames along the temporal dimension. SFEM: Sparse extraction is performed on features from different time steps to obtain key features within the time series. TFformer: The advanced temporal fusion attention mechanism is introduced to facilitate deep interactions between the current and historical frames. We validated the effectiveness of STFNet on the nuScenes dataset, achieving 71.8% NuScenes detection score (NdS) and 67.0% mean average precision (mAP). Compared to the benchmark method, our method improves 1.6% NdS and 1.5% mAP. Extensive experiments demonstrate that STFNet significantly outperforms most existing methods, highlighting the superiority and generalizability of our approach.

关键词： Feature extraction Three-dimensional displays Point cloud compression object detection Laser radar History Sensors Proposals Heating systems Fuses 3d object detection autonomous vehicle LidAR point cloud

来源：评论

学校读者我要写书评

暂无评论

Multimodal 3d object detection Based on Sparse Interaction in Internet of Vehicles

引用

IEEE TRANSACTIONS ON VEHICULAR TECHNOLOGY 2025年第2期74卷 2174-2186页

作者： Li, Hui Ge, Tongao Bai, Keqiang Nie, Gaofeng Xu, Lingwei Ai, Xiaoxue Cao, Song Qingdao Univ Sci & Technol Dept Informat Sci & Technol Qingdao 266061 Peoples R China Beijing Univ Posts & Telecommun State key Lab Networking & Switching Technol Beijing 100876 Peoples R China Guangxi Minzu Univ Guangxi Key Lab Hybrid Computat & IC Design Anal Nanning 530006 Peoples R China Southwest Univ Sci & Technol Sch Informat Engn Mianyang 621010 Sichuan Peoples R China

Combining the Internet of Vehicles with autonomous driving visual perception can enhance vehicle intelligence. Vehicles use the 3d object detection algorithm to perceive their surroundings and share detection results with other vehicles using the internet of vehicles technology, improving the efficiency of intelligent transportation systems. Multimodal information fusion of LidAR and cameras can improve the performance of 3d object detection. However, the different modality information is inhomogeneous, multimodal 3d object detection still has challenges, such as difficult semantic alignment of modal elements and inadequate fusion. To mitigate these challenges, we first propose the sparse interaction with centroid query (SICQ) for voxel-level features from different modalities, which aligns different modal semantic information through more precise and fine-grained interaction. Then, we propose the dense fusion with multi-scale masked attention (dFMMA), using multi-scale feature masks from bird's-eye-view (BEV)-level multimodal features to improve performance for small object feature perception. Finally, we propose the multimodal grid encoder with positional information (MGEPI), through positional information implicitly guiding and the transformer-based attention mechanism for grid-level features, improves the perception of detection scene context spatial information and enhances the robustness of the algorithm. Additionally, this paper performs comprehensive experiments on the popular KITTI dataset and demonstrates that our algorithm has superior 3d object detection performance.

关键词： Three-dimensional displays Feature extraction object detection Laser radar Point cloud compression Cameras Semantics Vehicle technology multimodal fusion 3d object detection Internet of Vehicles

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：