检索结果-内蒙古大学图书馆

2023 China Automation Congress, CAC 2023

作者： Huo, Jiaxin Sun, Lihang Liu, Jianyi Institute of Artificial Intelligence and Robotics Xi'an Jiaotong University Xi'an China

ISBN: (纸本)9798350303759

Current Scene Change Detection(SCD) methods are widely used in various subject areas, with detection granularity mostly limited to pixel-level. However, for certain practical applications such as garbage detection and traffic monitoring, the overall changes of object-level instances are more concerned so that fine-grained results may not be necessary, incurring excessive computational redundancy and insufficient real-time performance. To address the issue, we propose a one-stage object-level change detection framework named Siamese Center-Based Detector with Transformer and Feature Fusion (SCTF-Det), aiming at using less computing resources while still obtaining object-level change information, such as appearance or disappearance of objects. We adopt Siamese Vision Transformer to efficiently capture global semantic features and design differential feature fusion and multi-scale fusion to better fuse the features coming from image pairs. Instead of using a segmentation head like most SCD methods, we use a detection head to capture changed objects or regions. Moreover, we introduce a gating mechanism in image pairs and automatically mark the bounding box on the corresponding 'Appear' change region. The experiments are conducted on VL-CMU-CD and CDNet2014 datasets, with Fl scores of 78.6% and 83.6% respectively. Our SCTF-Det substantially improves inference speed by 3-5 times compared to the existing methods. © 2023 IEEE.

关键词： Semantics

来源：评论

学校读者我要写书评

暂无评论

A Novel Map Representation for Automatic Construction of Topological Map and Lane-Level Route Planning

A Novel Map Representation for Automatic Construction of Top...

引用

2023 China Automation Congress, CAC 2023

作者： Sun, Jian Xue, Jianru Li, Gengxin Liu, Jingcheng Institute of Artificial Intelligence and Robotics Xi' An Jiaotong University Xi'an China

ISBN: (纸本)9798350303759

Traditional road-level topological maps cannot meet the demand for high precision navigation services required by human drivers and autonomous vehicles. Meanwhile, current enhanced lane-level topological maps with redundant details pose practical challenges for automatic map construction and efficient lane-level route planning. To address these challenges, this paper proposes a novel topological map representation. Based on this representation, a hierarchical mapping framework is established, which includes the extraction of road primitives, generation of directed graph, and automatic construction of lane-level topological map. The constructed topological map enables efficient lane-level route planning by the shortest path search algorithm with having the minimal nodes of topological graph. The proposed mapping framework and lane-level route planning have been extensively tested and evaluated on two real urban traffic scenarios, demonstrating the feasibility and effectiveness of this novel topological map representation. © 2023 IEEE.

关键词： Autonomous vehicles

来源：评论

学校读者我要写书评

暂无评论

Learning to Embed Seen/Unseen Compositions based on Graph Networks

Learning to Embed Seen/Unseen Compositions based on Graph Ne...

引用

2023 China Automation Congress, CAC 2023

作者： Jiang, Dongyao Chen, Hui Ma, Yongqiang Jing, Haodong Zheng, Nanning Institute of Artificial Intelligence and Robotics Xi'An Jiaotong University Xi'an China

ISBN: (纸本)9798350303759

Composability allows known concepts to form newer and more complex ones. This coupling process is the research interests of Compositional Zero-Shot Learning (CZSL). The goal can be described as building a classifier for unknown compositions in the testing set based on known attribute primitives (e.g., old, cute) and object primitives (e.g., cats, cars) in the training set. There are many challenges in this process. For example, the same attribute primitive behaves significantly distinct on different objects. Common CZSL methods introduce auxiliary classification information into the model by using the pretrained model or external knowledge base, but the distribution of introduced auxiliary information is usually inconsistent with the distribution of the class information contained in training set itself, resulting in the model's misunderstanding of combined features. In view of this deficiency, we proposed a novel Compositional Graph Convolutional Network model, which consists of two embedding networks for image and label text modal data respectively. With graph convolutional networks, we can eliminate potential differences in information distribution in the pre-training data. Besides, we add a quintuplet loss to cross-entropy loss to generate a smoother feature representation. The results on three benchmarks including MIT-States, UT-Zappos and C-GQA datasets show that the proposed model exceeds the seven state-of-the-art methods in terms of Area Under the Curve (AUC) and classification accuracy. This confirms that our method can generate more recognizable compositional features. © 2023 IEEE.

关键词： Zero-shot learning

来源：评论

学校读者我要写书评

暂无评论

U-YOLO: Improved YOLOv5 for Small Object Detection on UAV-Captured Images 1st

U-YOLO: Improved YOLOv5 for Small Object Detection on UAV-...

引用

1st International Conference on Cognitive Computation and Systems, ICCCS 2022

作者： Zhang, Guowei Chen, Xingyu Tan, Xun Zhang, Jiahao Lan, Xuguang Institute of Artificial Intelligence and Robotics Xi’an Jiaotong University Xi’an China

ISBN: (纸本)9789819927883

Small object detection on drone-captured images is a recently popular and challenging task. From the drone’s perspective, the object scale varies significantly, and tiny objects lack distinguishable appearance information in complex backgrounds, which leads to undesired effects. To solve the issue mentioned above, we propose U-YOLO based on the original YOLOv5 model. We first extend the multi-scale feature fusion network and add a detection head for tiny objects. Secondly, we integrate the convolutional block attention model (CBAM) in the detection head to focus on the critical region of the feature map. Lastly, an attention feature fusion module based on contextual information is designed to combine local and global contextual details of small objects and attention mechanisms to enhance the multi-scale feature fusion capability. Experiments on dataset VisDrone2021-DET show that U-YOLO not only improves the detection performance on drone-captured scenarios but also has a real-time detection speed. Compared to baseline model (YOLOv5s), the mAP result of U-YOLO is increased from 19.07 to 24.1%, and the detection speed is 39FPS. It provides a good balance between detection accuracy and speed, promoting the progress of small object detection algorithms on UAV platforms. © 2023, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.

关键词： Object detection

来源：评论

学校读者我要写书评

暂无评论

Atten-ganCV: An End-to-End Close-Coupled Image-Generating Cross-View Network

Atten-ganCV: An End-to-End Close-Coupled Image-Generating Cr...

引用

2023 China Automation Congress, CAC 2023

作者： Song, Wenjie Xue, Jianru Institute of Artificial Intelligence and Robotics Xi'An Jiaotong University Xi'an China

ISBN: (纸本)9798350303759

Cross-view geo-localization aims to match the query input ground-view image and the aerial-view images in the reference dataset one by one to determine the ground image's geographic location. This research is extremely challenging because the variation of the observation angle between cross-view images brings about great geometric appearance differences between image pairs. Nowadays, the introduction of generative networks into matching models has been shown to work well on the CVUSA (Cross-View USA) dataset, and the latest models clarify the paradigm of end-to-end generative cross-view image matching methods. However, this result relies on an assumption on the dataset: for all query input ground images, there must exist a reference aerial image that is exactly centered on the location of that image, which is clearly not consistent with real-world application scenarios;and the performance of state-of-the-art generative models degrades significantly when departing from this assumption of center alignment. To address this problem, this paper provides a generative model (atten-ganCV) for non-center-aligned datasets. This model feeds the query ground image directly into a generative adversarial network to obtain a generated aerial view image, where the generator atten-UNet innovatively introduces an attention mechanism. Then, model matches the synthesized image with the real aerial image in the reference dataset one by one, and finally obtains the matching result with the highest similarity, thus determining the geographic location of the query input. The model is tested on both the center-aligned CVUSA dataset and the non-center-aligned VIGOR (Cross-view Image Geo-localization beyond One-to-one Retrieval) dataset. In the VIGOR dataset, this model achieves approximately the same accuracy as the state-of-the-art model with 3 times the inference speed. © 2023 IEEE.

关键词： Antennas

来源：评论

学校读者我要写书评

暂无评论

Learning Object Relation Graph and Goal-Aware Policy For Visual Room Rearrangement

Learning Object Relation Graph and Goal-Aware Policy For Vis...

引用

2023 China Automation Congress, CAC 2023

作者： Wang, Beibei Wang, Xiaohan Liu, Yuehu College of Artificial Intelligence Xi'an Jiaotong University Xi'an China Institute of Artificial Intelligence and Robotics College of Artificial Intelligence Xi'an Jiaotong University Xi'an China

ISBN: (纸本)9798350303759

Embodied AI, where agents accomplish specific tasks through interaction with their surrounding environment, is attracting attention in the community. As a more comprehensive and practical embodied task, visual room rearrangement aims to rearrange the initially misplaced objects in a room to the target configuration. In this task, the misplaced targets are mostly small objects that are difficult to locate, and there is no indication of which object needs to be rearranged in sensory input. Direct search for small and unknown targets poses a challenge. Previous works tend to directly incorporate sensory input into a reward-based framework. These methods do not learn the relationships between objects and infer which objects need to be rearranged, resulting in low performance. Understanding the scene structure and reasoning task goals can enable agents to rearrange targets more efficiently. In this paper, we propose a receptacle-target relation graph, which considers the spatial relationship between target objects and the receptacle objects appearing around them, enabling agents to understand the environment. Furthermore, we also propose a target reasoning network to explicitly predict the target object. Finally, our goal-aware policy can plan a series of actions based on their understanding of the task goal. Extensive compmethodarative and ablation experiments were conducted on the AI2Thor platform, demonstrating the effectiveness of our method. © 2023 IEEE.

关键词： Computer vision

来源：评论

学校读者我要写书评

暂无评论

Safe Multi-Agent Reinforcement Learning for Behavior-Based Cooperative Navigation

引用

IEEE robotics and Automation Letters 2025年第6期10卷 6256-6263页

作者： Dawood, Murad Pan, Sicong Dengler, Nils Zhou, Siqi Schoellig, Angela P. Bennewitz, Maren University of Bonn Humanoid Robots Lab Bonn53113 Germany Lamarr Institute for Machine Learning and Artificial Intelligence The Center for Robotics Bonn53113 Germany Technical University of Munich Learning Systems and Robotics Lab Munchen80333 Germany

In this letter, we address the problem of behavior-based cooperative navigation of mobile robots usingsafe multi-agent reinforcement learning (MARL). Our work is the first to focus on cooperative navigation without individual reference targets for the robots, using a single target for the formation's centroid. This eliminates the complexities involved in having several path planners to control a team of robots. To ensure safety, our MARL framework uses model predictive control (MPC) to prevent actions that could lead to collisions during training and execution. We demonstrate the effectiveness of our method in simulation and on real robots, achieving safe behavior-based cooperative navigation without using individual reference targets, with zero collisions, and faster target reaching compared to baselines. Finally, we study the impact of MPC safety filters on the learning process, revealing that we achieve faster convergence during training and we show that our approach can be safely deployed on real robots, even during early stages of the training. © 2025 IEEE.

关键词： Adversarial machine learning

来源：评论

学校读者我要写书评

暂无评论

DO GENERATED DATA ALWAYS HELP CONTRASTIVE LEARNING? 12

DO GENERATED DATA ALWAYS HELP CONTRASTIVE LEARNING?

引用

12th International Conference on Learning Representations, ICLR 2024

作者： Wang, Yifei Zhang, Jizhe Wang, Yisen School of Mathematical Sciences Peking University China Institute of Artificial Intelligence and Robotics Xi'an Jiaotong University China National Key Lab of General Artificial Intelligence School of Intelligence Science and Technology Peking University China Institute for Artificial Intelligence Peking University China

Contrastive Learning (CL) has emerged as one of the most successful paradigms for unsupervised visual representation learning, yet it often depends on intensive manual data augmentations. With the rise of generative models, especially diffusion models, the ability to generate realistic images close to the real data distribution has been well recognized. These generated high-equality images have been successfully applied to enhance contrastive representation learning, a technique termed "data inflation". However, we find that the generated data (even from a good diffusion model like DDPM) may sometimes even harm contrastive learning. We investigate the causes behind this failure from the perspective of both data inflation and data augmentation. For the first time, we reveal the complementary roles that stronger data inflation should be accompanied by weaker augmentations, and vice versa. We also provide rigorous theoretical explanations for these phenomena via deriving its generalization bounds under data inflation. Drawing from these insights, we propose Adaptive Inflation (AdaInf), a purely data-centric strategy without introducing any extra computation cost. On benchmark datasets, AdaInf can bring significant improvements for various contrastive learning methods. Notably, without using external data, AdaInf obtains 94.70% linear accuracy on CIFAR-10 with SimCLR, setting a new record that surpasses many sophisticated methods. Code is available at https://***/PKU-ML/adainf. © 2024 12th International Conference on Learning Representations, ICLR 2024. All rights reserved.

关键词： Image enhancement

来源：评论

学校读者我要写书评

暂无评论

Event-enhanced synthetic aperture imaging

引用

Science China(Information Sciences) 2025年第3期68卷 388-390页

作者： Siqi LI Shaoyi DU Jun-Hai YONG Yue GAO Beijing National Research Center for Information Science and Technology School of SoftwareTsinghua University Institute for Brain and Cognitive Sciences Beijing Laboratory of Brain and Cognitive IntelligenceTsinghua University National Key Laboratory of Human-Machine Hybrid Augmented Intelligence National Engineering Research Center for Visual Information and ApplicationsInstitute of Artificial Intelligence and Robotics Xi'an Jiaotong University

Synthetic aperture imaging(SAI) methods aim to see through dense occlusions and reconstruct the target scene behind occlusions. Traditional frame-based SAI methods,e.g., DeOccNet [1], take the occluded light field ima...

关键词：

来源：评论

学校读者我要写书评

暂无评论

A Sparsity-Aware Autonomous Path Planning Accelerator with Algorithm-Architecture Co-Design 24

A Sparsity-Aware Autonomous Path Planning Accelerator with A...

引用

43rd International Conference on Computer-Aided Design, ICCAD 2024

作者： Zhang, Yanjun Niu, Xiaoyu Zhang, Yifan Tian, Hongzheng Yu, Bo Liu, Shaoshan Huang, Sitao Beijing Institute of Technology China University of California Irvine United States Shenzhen Institute of Artificial Intelligence and Robotics for Society China

ISBN: (纸本)9798400710773

Path planning is a critical task in autonomous driving systems that is most susceptible to real-time constraints but often demands computationally intensive mathematical solvers, two contradictory goals. This conflict makes the computing of path planning a paramount challenge. At the heart of most path planners is the quadratic programming (QP) solver, which places excessive demands on the CPU in real-world autonomous driving applications. In this paper, we present an FPGA-based acceleration framework for path planning problems. Our approach leverages an operator splitting solver for quadratic programs (OSQP) and employs the preconditioned conjugate gradient (PCG) method for solving linear systems, which are customized to be more hardware-friendly than prior works. Specific memory management and parallel processing were tailored to the matrix pattern, and the incorporation of pipelining was executed to enhance throughput and execution speed. Our FPGA-based implementation achieves state-of-the-art performance against existing works, including an average 1.98× speedup compared with the state-of-the-art QP solver on Intel i7-11800H CPU, 3.90× speedup over an ARM Cortex-A57 embedded CPU, and 12.3× speedup over an NVIDIA RTX 3090 GPU. © 2024 Copyright is held by the owner/author(s).

关键词： Quadratic programming

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：