检索结果-内蒙古大学图书馆

Multi-scale inputs and context-aware aggregation network for stereo matching

MULTIMEDIA TOOLS AND APPLICATIONS 2024年第30期83卷 75171-75194页

作者： Shi, Liqing Xiong, Taiping Cui, Gengshen Pan, Minghua Cheng, Nuo Wu, Xiangjie Guilin Univ Elect Technol Guangxi Key Lab Image & Graph Intelligent Proc Guilin 541004 Peoples R China Guilin Univ Elect Technol Sch Comp Sci & Informat Secur Guilin 541004 Peoples R China Guilin Univ Elect Technol Nanning Res Inst Nanning 530000 Peoples R China

Despite the significant progress made in deep learning-based stereo matching, the accuracy of these methods significantly decreases when faced with challenges such as occlusions, reflections, textureless areas, and scale variations. In this paper, we propose MSCANet, a novel stereo matching network that integrates multi-scale inputs and context-aware aggregation ability. MSCANet effectively integrates rich multi-scale feature information and exhibits context-aware capability, thereby enabling it to achieve superior performance. Firstly, a multi-scale aware fusion module is designed to efficiently incorporate more comprehensive global context features at different scales, which allows the model to enhance its ability to generalize across images of varying scales. Secondly, a novel V-shaped encoder/decoder module is developed to effectively exploit the rich feature information. In the encoding stage, a 3D squeeze-and-excitation block is introduced to facilitate adaptively recalibration of learned feature maps. This block effectively suppresses irrelevant features while enhancing useful features, which improved efficiency and accuracy in disparity prediction. Additionally, a 3D context-aware decode block is designed to effectively utilize global context features to restore the original image structure during the decoding stage. Moreover, the high-level feature maps can be employed to augment low-level feature maps by incorporating more detailed information to avoid the side effects caused by the loss of information during the encoding process. Extensive ablation experiments and comparative experiments were conducted on Scene Flow dataset, KITTI2012 and KITTI2015 datasets to validate the effectiveness of each proposed module. The experimental results demonstrate MSCANet achieves competitive performance and offers a more straightforward and efficient model design, as well as faster inference speed.

关键词： Multi-scale feature fusion Context-aware capability 3D squeeze-and-excitation Stereo matching Binocular vision

来源：评论

学校读者我要写书评

暂无评论

A Consumer-Oriented image Transformation Scheme With a Secret key for Privacy Protection

引用

IEEE SIGNAL procESSING LETTERS 2024年 31卷 2280-2284页

作者： Zhao, Ruoyu Wen, Wenying Zhu, Youwen Lan, Rushi Nanjing Univ Aeronaut & Astronaut Coll Comp Sci & Technol Nanjing 211106 Peoples R China Jiangxi Univ Finance & Econ Sch Informat Technol Nanchang 330013 Peoples R China Guilin Univ Elect Technol Guangxi Key Lab Image & Graph Intelligent Proc Guilin 541004 Peoples R China

images in electronic devices may pose privacy threats since they can capture sensitive information about consumers. Meanwhile, face recognition (FR) systems are widely used, exacerbating these concerns. Some learning-based schemes have been proposed to protect the privacy of facial images. However, consumers might encounter challenges in implementing them due to specific requirements related to computing power and professional background. The non-learning semantic adversarial perturbation schemes address the aforementioned issues, but they are either irreversible or necessitate additional storage space. To this end, we propose a consumer-oriented image transformation scheme that can prevent the recognition of images by FR systems. It offers a consumer-friendly method compared to schemes that need high-performance equipment and specialized knowledge. The proposed scheme is implemented by reversible embedding the key to coefficients during the JPEG compression. The proposed scheme can be operated in general computing devices by ordinary consumers. The experiments demonstrated that facial images can be protected by the proposed scheme.

关键词： Privacy Signal processing algorithms Semantics Perturbation methods Vectors Transform coding Protection Consumer friendliness facial image image transformation JPEG compatibility privacy preservation

来源：评论

学校读者我要写书评

暂无评论

A survey of high utility sequential patterns mining methods

引用

JOURNAL OF intelligent & FUZZY SYSTEMS 2023年第5期45卷 8049-8077页

作者： Zhang, Ruihua Han, Meng He, Feifei Meng, Fanxing Li, Chunpeng North Minzu Univ Sch Comp Sci & Engn Yinchuan Ningxia Peoples R China State Ethn Affairs Commiss Key Lab Intelligent Proc Image & Graph Yinchuan Ningxia Peoples R China

In recent years, there has been an increasing demand for high utility sequential pattern (HUSP) mining. Different from high utility itemset mining, the "combinatorial explosion" problem of sequence data makes it more challenging. This survey aims to provide a general, comprehensive, and structured overview of the state-of-the-art methods of HUSP from a novel perspective. Firstly, from the perspective of serial and parallel, the data structure used by the mining methods are illustrated and the pros and cons of the algorithms are summarized. In order to protect data privacy, many HUSP hiding algorithms have been proposed, which are classified into array-based, chain-based and matrix-based algorithms according to the key technologies. The hidden strategies and evaluation metrics adopted by the algorithms are summarized. Next, a taxonomy of the most common and the state-of-the-art approaches for incremental mining algorithms is presented, including tree-based and projection-based. In order to deal with the latest sequence in the data stream, the existing algorithms often use the window model to update dynamically, and the algorithms are divided into methods based on sliding windows and landmark windows for analysis. Afterwards, a summary of derived high utility sequential pattern is presented. Finally, aiming at the deficiencies of the existing HUSP research, the next work that the author plans to do is given.

关键词： Survey high utility sequential patterns incremental data data streams hidden patterns

来源：评论

学校读者我要写书评

暂无评论

PLM-YOLOv5: Improved YOLOv5 for Pig Detection in Livestock Monitoring

PLM-YOLOv5: Improved YOLOv5 for Pig Detection in Livestock M...

引用

2024 International Conference on intelligent Robotics and Automatic Control

作者： Peng, Ni Li, Fang Luo, Xiaonan Guilin Univ Elect Technol Sch Comp Sci & Informat Secur Guilin 541004 Peoples R China Ctr Appl Math Guangxi GUET Guilin 541004 Peoples R China Guangxi Key Lab Image & Graph Intelligent Proc Guilin 541004 Peoples R China

ISBN: (纸本)9798350389814;9798350389807

With the rapid development of smart agriculture, the demand for pig monitoring has been increasing, particularly in disease prevention and behavioral analysis. Real-time monitoring of pig behavior is essential for improving farming efficiency and ensuring animal welfare. However, challenges such as pigs' varying body sizes, movement, and complex environments make accurate detection and tracking difficult. To address these issues, this paper proposes an improved YOLOv5 model, named PLM-YOLOv5, specifically designed for pig detection in livestock monitoring. First, we add an additional detection head to enhance the model's capability to detect smaller pigs effectively. Next, we integrate the Convolutional Block Attention Module (CBAM) to enhance feature extraction, particularly in handling complex backgrounds. Finally, we utilize a self-collected dataset of over 1,400 images for model training and validation, ensuring the model adapts to real-world livestock environments. Experimental results show that PLM-YOLOv5 achieves significant improvements in pig detection tasks, with the mean Average Precision (mAP) increasing by approximately 9% compared to the original YOLOv5. Additionally, training and validation loss curves demonstrate the model's stability and convergence. These results confirm the effectiveness and reliability of PLM-YOLOv5 for real-world applications, offering a robust solution for smart agriculture.

关键词： YOLOv5 Livestock Monitoring Smart Agriculture CBAM Pig Detection

来源：评论

学校读者我要写书评

暂无评论

A survey of intelligent optimization algorithms for solving satisfiability problems

引用

JOURNAL OF intelligent & FUZZY SYSTEMS 2023年第1期45卷 445-461页

作者： Yang, Lan Wang, Xiaofeng Ding, Hongsheng Yang, Yi Zhao, Xingyu Pang, Lichao North Minzu Univ Sch Comp Sci & Engn Yinchuan Peoples R China North Minzu Univ Key Lab Image & Graph Intelligent Proc State Ethn Affairs Commiss Yinchuan Peoples R China

Constraint satisfaction problems have a wide range of applications in areas such as basic computer theory research and artificial intelligence, and many major studies in industry are not solved directly, but converted into instances of satisfiability problems for solution. Therefore, the solution of the satisfiability problem is a central problem in many important areas in the future. A large number of solution algorithms for this problem are mainly based on completeness algorithms and heuristic algorithms. intelligent optimization algorithms with heuristic policies run significantly more efficiently on large-scale instances compared to completeness algorithms. This paper compares the principles, implementation steps, and applications of several major intelligent optimization algorithms in satisfiability problems, analyzes the characteristics of these algorithms, and focuses on the performance in solving satisfiability problems under different constraints. In terms of algorithms, evolutionary algorithms and swarm intelligence algorithms are introduced;in terms of applications, the solution to the satisfiability problem is studied. At the same time, the performance of the listed intelligent optimization algorithms in applications is analyzed in detail in terms of the direction of improvement of the algorithms, advantages and disadvantages and comparison algorithms, respectively, and the future application of intelligent optimization algorithms in satisfiability problems is prospected.

关键词： Constraint satisfaction problem satisfiability problem completeness algorithm heuristic algorithm intelligent optimization algorithms

来源：评论

学校读者我要写书评

暂无评论

Efficient Pose Estimation via a Lightweight Single-Branch Pose Distillation Network

引用

IEEE SENSORS JOURNAL 2023年第22期23卷 27709-27719页

作者： Zhang, Shihao Qiang, Baohua Yang, Xianyi Zhou, Mingliang Chen, Ruidong Chen, Lirui Guilin Univ Elect Technol Guangxi Key Lab Image & Graph Intelligent Proc Guilin 541004 Peoples R China Luohe Vocat & Tech Coll Sch Informat Engn Luohe 462000 Peoples R China Chongqing Univ Sch Comp Sci Chongqing 400044 Peoples R China

Accurate lightweight (LW) pose estimation is still a challenging task influenced by different human poses and various complex backgrounds in 2-D human images. To address the above problems, we propose a lightweight single-branch pose distillation network, termed LSPD, which is a lightweight powerful fully convolutional pose network that can be executed quickly with a low computational cost for accurate pose estimation. First, we introduced an efficient end-to-end pose distillation sequence framework, which utilizes a small number of lightweight and strong pose estimation stages to effectively transfer the pose knowledge of our teacher model. Second, we constructed a compact and strong pose estimation stage that uses a type of lightweight multiscale residual block to enhance the image features and the image-dependent spatial features representation ability of the model. At the same time, it reduces the computational cost. Finally, when training is complete, we used the backbone network and the first student stage as the simple architecture to deploy. Extensive experiments demonstrated that the proposed method obtains excellent performance with high accuracy and low model parameters.

关键词： Efficient pose estimation end-to-end low model parameters pose distillation

来源：评论

学校读者我要写书评

暂无评论

Knowledge Distillation for Lightweight 2D Single-Person Pose Estimation

引用

JOURNAL OF CIRCUITS SYSTEMS AND COMPUTERS 2023年第3期32卷 2350050-2350050页

作者： Zhang, Shihao Qiang, Baohua Yang, Xianyi Zhou, Mingliang Chen, Ruidong Guilin Univ Elect Technol Guangxi Key Lab Image & Graph Intelligent Proc Guilin 541004 Peoples R China Luohe Vocat Technol Coll Sch Informat Engn Luohe 462000 Peoples R China Chongqing Univ Sch Comp Sci Chongqing 400044 Peoples R China

The current state-of-the-art single-person pose estimation methods require heavily parameterized models for accurate predictions. A promising technique to achieve accurate yet lightweight pose estimation is knowledge distillation. However, existing pose knowledge distillation methods rely on the most common large basic building blocks and a complex multi-branch architecture. In this study, we propose a Single-branch Lightweight Knowledge Distillation method to increase pose distillation efficiency for 2D Single-person pose estimation, termed SLKD2S. First, we design a novel single-branch pose knowledge distillation framework, which is composed of connected lightweight pose estimation stages. Second, we utilize a special pose distillation loss based on the joint confidence map. Finally, we only keep the initial stage and the first refinement stage to achieve a good performance. Extensive experiments on two standard benchmark datasets show the superiority of the proposed SLKD2S in terms of cost and accuracy, and the average detection accuracies are increased by 1.43% and 2.74% compared with the top-performing pose distillation method, respectively.

关键词： Single-person pose estimation pose distillation single branch refinement stage

来源：评论

学校读者我要写书评

暂无评论

Using Frequency Attention to Make Adversarial Patch Powerful Against Person Detector

引用

IEEE ACCESS 2023年 11卷 27217-27225页

作者： Lei, Xiaochun Cai, Xiang Lu, Chang Jiang, Zetao Gong, Zhaoting Lu, Linjun Guilin Univ Elect Technol GUET Sch Comp Sci & Informat Secur Guilin 541004 Peoples R China Guangxi Key Lab Image & Graphic Intelligent Proc Guilin 541004 Peoples R China

Deep neural networks (DNNs) are vulnerable to adversarial attacks. In particular, object detectors may be attacked by applying a particular adversarial patch to the image. However, because the patch will lose information during the shrinking process, most existing approaches that employ adversarial patches to attack object detectors would diminish the attack success rate on small and medium targets. This paper proposes a Frequency Module(FRAN), a frequency-domain attention module for guiding patch generation to solve this problem. We use FRAN to allow the patched attack to rely on the low-frequency signals which are lost less in the shrinking process. This is the first study to introduce frequency domain attention to optimize the attack capabilities of adversarial patches. Our method increases the attack success rates of small and medium targets by 4.18% and 3.89%, respectively, over the state-of-the-art attack method for fooling the human detector while assaulting YOLOv3 without reducing the attack success rate of big targets.

关键词： Detectors Frequency-domain analysis Object detection Training Transform coding Deep learning Transformers Adversarial machine learning Adversarial attack frequency attention object detection deep learning

来源：评论

学校读者我要写书评

暂无评论

A Weakly Supervised Semantic Segmentation Method on Lung Adenocarcinoma Histopathology images 19th

A Weakly Supervised Semantic Segmentation Method on Lung Ade...

引用

19th International Conference on Advanced intelligent Computing Technology and Applications (ICIC)

作者： Lan, Xiaobin Mei, Jiaming Lin, Ruohan Chen, Jiahao Zhang, Yanju Guilin Univ Elect Technol Guangxi Key Lab Image & Graph Intelligent Proc Guilin Guangxi Peoples R China Huaqiao Univ Coll Comp Sci & Technol Xiamen Peoples R China Huaqiao Univ Xiamen Key Lab CVPR Xiamen Peoples R China

ISBN: (纸本)9789819947416;9789819947423

Accurately distinguishing different tissues is crucial for pathologists to determine the type and degree of lesions. The methods based on class activation maps (CAM) are popular in current research due to their low annotation cost. However, this method has two main drawbacks: (1) as the training deepens, the class activation maps become increasingly concentrated in the most discriminative region and cannot fully cover the object;(2) different tissues may be activated at the same pixel because the binary cross-entropy (BCE) calculates classification loss independently. To alleviate these issues, we propose a new method that combines channel attention (CA) and progressive dropout attention (PDA) to expand the class activation maps and enhance the activation of important features. In addition, after the network converges, we train an additional fully-connected layer and use softmax cross-entropy (SCE) to fine-tune the model parameters and reduce the activation of unrelated tissues. We evaluate our proposed model on the recently released LUAD-HistoSeg dataset. The results illustrate that our model achieved a 76.53% MIoU and outperform the most advanced methods.

关键词： tissue segmentation weakly supervised semantic segmentation attention mechanism class re-activation mapping

来源：评论

学校读者我要写书评

暂无评论

Event-Triggered Control of Switched Two-Time-Scale Systems With Asynchronous Switching

引用

IEEE CONTROL SYSTEMS LETTERS 2024年 8卷 2075-2080页

作者： Zeng, Ze-Hong Wang, Yan-Wu Liu, Xiao-Kang Yang, Wu Huazhong Univ Sci & Technol Sch Artificial Intelligence & Automat Key Lab Image Proc & Intelligent Control Wuhan 430074 Peoples R China Guangxi Univ Sch Elect Engn Nanning 530004 Peoples R China

This letter addresses the event-triggered control problem for uncertain switched two-time-scale systems (TTSSs) experiencing asynchronous switching. Due to ill-conditioning issues, methods developed for switched single-time-scale systems cannot be directly applied to switched TTSSs. Additionally, event-triggered results for non-switched TTSSs are inapplicable due to the complexities introduced by switching behavior and asynchronism. To address these challenges, we utilize a Lyapunov function dependent on the singular perturbation parameter and the controller mode to establish conditions sufficient to ensure the exponential stability of the closed-loop system with average dwell time switching. These conditions facilitate the joint design of a mode-dependent event-triggered mechanism and the control gains while mitigating the risk of Zeno behavior. Finally, a numerical example is provided to demonstrate the effectiveness and benefits of the proposed theorem.

关键词： Switches Control systems Switched systems Symmetric matrices Synchronization Switching circuits Perturbation methods LMIs Lyapunov methods sampled-data control

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：