检索结果-内蒙古大学图书馆

DFS-Net: StyleGAN2-based dual feature separation for face De-Morphing

Multimedia Tools and Applications 2024年 1-23页

作者： Long, Ming Chen, Song Zhang, Le-Bing Peng, Fei Zhang, Dengyong School of Computer and Communication Engineering Changsha University of Science and Technology Changsha410114 China School of Computer and Artificial Intelligence Huaihua University Huaihua418000 China Institute of Artificial Intelligence and Blockchain Guangzhou University Guangzhou510006 China

Face morphing attacks pose a significant threat to face recognition systems. In order to solve this problem, several methods for detecting these attacks have been proposed. However, the restoration of the accomplice’s face image remains in its nascent developmental stage. In this paper, we introduce a novel network architecture termed DFS-Net, which leverages double feature spaces (latent code and feature tensor) and a dual-feature separation (DFS) network. DFS-Net is built upon the StyleGAN2 generator. By utilizing the feature vector and feature tensor extracted by the encoder, distinct separation networks are designed. This design facilitates effective feature separation, enabling DFS-Net to separate the identity features of accomplices. The incorporation of the feature tensor enhances the precision in extracting the identity features of the accomplice and thereby elevates the perceptual quality of the restoration image. Experimental results indicate that DFS-Net can effectively restore the accomplice’s face and outperforms previous works in terms of restoration accuracy and image visual quality. © The Author(s), under exclusive licence to Springer science+Business Media, LLC, part of Springer Nature 2024.

关键词： Network architecture

来源：评论

学校读者我要写书评

暂无评论

Arc Filtering Method Based on Generative Adversarial Networks 5

Arc Filtering Method Based on Generative Adversarial Network...

引用

5th International Symposium on computer Engineering and Intelligent Communications, ISCEIC 2024

作者： Zhang, Maihao Chen, Deng School of Computer Science & Engineering Artificial Intelligence Wuhan Institute of Technology Wuhan China

ISBN: (纸本)9798331518677

Robot monitoring has been widely applied to observe the operation of welding robots, but the arc generated during the welding process causes severe overexposure in local areas of robot surveillance images, significantly deteriorating the monitoring effect of the welding robots. Due to the complex background structure of robot welding images, existing methods produce poor quality images after filtering out the arc from robot welding images. In this study, we propose an arc filtering model based on generative adversarial networks (GANs), primarily aimed at filtering out the arc in robot welding images. The well-trained arc filtering model can effectively eliminate the arc in robot welding images, generating high-quality non-arcing images. Experimental studies were conducted using an industrial robot welding image dataset we built autonomously and compared with state-of-the-art methods, demonstrating the superiority of our proposed model. © 2024 IEEE.

关键词： Generative adversarial networks

来源：评论

学校读者我要写书评

暂无评论

Defect Detection on Wire Based on YOLOv8 3

Defect Detection on Wire Based on YOLOv8

引用

3rd International Conference on artificial intelligence and computer Information technology, AICIT 2024

作者： Guo, Kai Yu, Baocheng School of Computer Science & Engineering Artificial Intelligence Wuhan Institute of Technology Wuhan China

ISBN: (纸本)9798350365443

The false detection rate and missed detection rate of wire defects using manual detection is unstable, when automatic winding machines are applied. Aiming at addressing the problem, a defect detection model based on YOLOv8 named Wire-YOLOv8 was proposed. First of all, SPD-Conv module is used to replace Conv module in the backbone to help extracting multi-scale features and obtain features from small defects, in this way, the ability to extract features is enhanced, detection accuracy is improved and miss detection rate drops. Furthermore, SimSPPF take the place of SPPF to reduce the size of computation and improve detection speed. Finally, WIoU is introduced to make loss function work better, optimizing localization ability of the model, providing high quality anchor and strong ability to detect defects. The experimental results show that the average accuracy (mAP@0.5) of the improved model Wire-YOLOv8 for the grease dirt, outer insulation defect and inner insulation defect of wire reaches 92.7%, 94.6% and 97.3%, which is 1.5%, 2.3% and 1.9% higher than the original model, respectively, and the detection speed reaches 211FPS, which can serve as a compelling solution to detect wire defects in industrial sites. © 2024 IEEE.

关键词： Wire

来源：评论

学校读者我要写书评

暂无评论

Construction of maritime traffic network based on DBSCAN 4

Construction of maritime traffic network based on DBSCAN

引用

2024 4th International Conference on artificial intelligence and Industrial technology Applications, AIITA 2024

作者： Wu, Hanyue School of Computer Science and Artificial Intelligence Wuhan University of Technology Wuhan430000 China

Trajectory data is essentially a sequence of spatial points ordered by timestamps, usually with some descriptive information in addition to basic spatiotemporal information. This paper investigates how publicly available Automatic Identification System (AIS) data can be used to analyze maritime traffic and transform it into directed graphs for estimating potential destination points of trajectories. In the maritime field, analyzing and modeling maritime traffic is crucial for vessel safety and efficiency, creating pathways with waypoints and segments, and further forming traffic networks. Our approach incorporates a detailed analysis of the distribution characteristics of different types of navigational points, leading to the adoption of tailored clustering parameters and methods. This differentiation allows for a nuanced understanding of stationary points at docks and anchorages, as well as navigational changes along shipping routes. Using the DBSCAN algorithm, we successfully cluster similar waypoints, considering cluster density and shape without needing to predefine the cluster count. © Published under licence by IOP Publishing Ltd.

关键词： Waterway transportation

来源：评论

学校读者我要写书评

暂无评论

A Memory-Guided Anomaly Detection Model with Contrastive Learning for Multivariate Time Series

引用

computers, Materials & Continua 2023年第11期77卷 1893-1910页

作者： Wei Zhang Ping He Ting Li Fan Yang Ying Liu School of Electronic and Information Engineering Hebei University of TechnologyTianjin300401China School of Artificial Intelligence Hebei University of TechnologyTianjin300401China School of Computer Science and Engineering University of Electronic Science and Technology of ChinaChengdu611731China

Some reconstruction-based anomaly detection models in multivariate time series have brought impressive performance advancements but suffer from weak generalization ability and a lack of anomaly *** limitations can result in the misjudgment of models,leading to a degradation in overall detection *** paper proposes a novel transformer-like anomaly detection model adopting a contrastive learning module and a memory block(CLME)to overcome the above *** contrastive learning module tailored for time series data can learn the contextual relationships to generate temporal fine-grained *** memory block can record normal patterns of these representations through the utilization of attention-based addressing and reintegration *** two modules together effectively alleviate the problem of ***,this paper introduces a fusion anomaly detection strategy that comprehensively takes into account the residual and feature *** a strategy can enlarge the discrepancies between normal and abnormal data,which is more conducive to anomaly *** proposed CLME model not only efficiently enhances the generalization performance but also improves the ability of anomaly *** validate the efficacy of the proposed approach,extensive experiments are conducted on well-established benchmark datasets,including SWaT,PSM,WADI,and *** results demonstrate outstanding performance,with F1 scores of 90.58%,94.83%,91.58%,and 91.75%,*** findings affirm the superiority of the CLME model over existing stateof-the-art anomaly detection methodologies in terms of its ability to detect anomalies within complex datasets accurately.

关键词： Anomaly detection multivariate time series contrastive learning memory network

来源：评论

学校读者我要写书评

暂无评论

EEDN: An Efficient Edge Detection Network for Aeroengine Blade Defect Segmentation 42

EEDN: An Efficient Edge Detection Network for Aeroengine Bla...

引用

42nd Chinese Control Conference, CCC 2023

作者： Wang, Chuhan Chen, Haiyong Hebei University of Technology School of Artificial Intelligence and Data Science China

ISBN: (纸本)9789887581543

A fast and accurate defect segmentation model is crucial for aeroengine blade inspection. However, most of the defects on the surface of aeroengine blades are micron-sized, which makes the defect features easily lost. In addition, the scales of defects on the surface of aeroengine blades vary widely. In this paper, we propose an Efficient Edge Detection Network (EEDN) for Aeroengine Blade Defect Segmentation. We design a low-complexity network backbone to learn feature information using depthwise separable convolutions efficiently. Moreover, in the decoding structure, we propose a Multiscale Feature Enhanced Attention (MFEA) to improve the multiscale expressiveness of the network and capture the long-range channel information. Experiments have proved that EEDN has achieved 0.859ODS and 97FPS on our dataset, which is superior to existing models. © 2023 Technical Committee on Control Theory, Chinese Association of Automation.

关键词： Edge detection

来源：评论

学校读者我要写书评

暂无评论

BACFuse: Toward Noise-Resistant BAC Detection Based on Multimodal Fusion on Smartphone

引用

IEEE Internet of Things Journal 2025年第11期12卷 16410-16423页

作者： Wu, Yuan Zhao, Gaorong Yang, Zheng Feng, Yong Wang, Yongmei Ding, Lei Chen, Jia Zhang, Jian Chen, Yanjiao Wuhan Textile University School of Computer Science and Artificial Intelligence Engineering Research Center of Hubei Province for Clothing Information Wuhan430200 China Wuhan University School of Computer Science Wuhan230036 China Kunming University of Science and Technology School of Information Engineering and Automation Kunming650500 China Anhui Agricultural University School of Information and Artificial Intelligence Hefei230036 China Zhejiang University College of Electrical Engineering Hangzhou310027 China

Drunk-driving is an important factor causing road traffic accidents and deaths, which deserves a lot of research. However, most current methods for detecting drunk-driving depend on customized hardware or require users’ active participation, making it impractical to monitor blood alcohol content (BAC) during driving. This article introduces BACFuse, a device-free, contactless, and noninvasive system utilizing smartphone in driving environments, which achieves relatively high accuracy in drunk-driving monitoring by integrating various voice sensing modalities. BACFuse first captures vocal cord vibration from ultrasonic signals, then records voice commands from audio signals. BACFuse combines the ultrasonic signals with audio signals and effectively detects drunk-driving and BAC. A key enabler lies in our modeling of latent interaction between acoustic and ultrasonic signals to mitigate ambient noise, realizing noise-resistant drunk-driving detection. Additionally, we propose an effective modules within the co-attention method to fuse the multimodal signals, further enhancing the accuracy of drunk-driving detection. We conduct extensive experiments to evaluate BACFuse’s performance on 20 participants in safe laboratory experiments. The results demonstrate that our system achieves BAC measurement with an MAE of 2.13 mg/dl, showing promise for future in-car driving management paradigms. © 2014 IEEE.

关键词： Ultrasonic devices

来源：评论

学校读者我要写书评

暂无评论

Decomposition-driven Multi-level Feature Fusion Framework for Infrared and Visible Image Fusion 5

Decomposition-driven Multi-level Feature Fusion Framework fo...

引用

5th International Symposium on computer Engineering and Intelligent Communications, ISCEIC 2024

作者： Wang, Haoming Gao, Shu School of Computer Science and Artificial Intelligence Wuhan University of Technology Hubei WuHan China

ISBN: (纸本)9798331518677

This paper proposes a novel infrared and visible image fusion method based on the feature decomposition and the Multi-level Feature Fusion. Firstly, modal features are initially extracted and decomposed into unique features with disparate foreground information and common features with similar background information to better discern the background to be maintained and the foreground to be emphasized. Then, the multi-level features are extracted and fused, thus the target image contains detailed and semantic information in low-level and high-level features respectively. Finally, the fused multi-level feature is reconstructed to the target image. Experiments on public datasets M3FD and MSRS show that the proposed method is better than other latest methods. © 2024 IEEE.

关键词： Image fusion

来源：评论

学校读者我要写书评

暂无评论

A Rearrangement and Restore-Based Mixer Model for Target-Oriented Multimodal Sentiment Classification

IEEE Transactions on Artificial Intelligence

引用

IEEE Transactions on artificial intelligence 2024年第6期5卷 3109-3119页

作者： Jia, Li Ma, Tinghuai Rong, Huan Sheng, Victor S. Huang, Xuejian Xie, Xintong Nanjing University of Information Science and Technology School of Computer Science Nanjing Jiangsu210044 China Nanjing University of Information Science and Technology School of Computer and Software Nanjing Jiangsu210044 China Jiangsu Ocean University School of Computer Engineering Lianyungang Jiangsu222005 China Nanjing University of Information Science and Technology School of Artificial Intelligence Nanjing Jiangsu210044 China Texas Tech University Department of Computer Science LubbockTX79409 United States

With the development of fine-grained multimodal sentiment analysis tasks, target-oriented multimodal sentiment (TMSC) analysis has received more attention, which aims to classify the sentiment of target with the help of textual and associated image features. Existing methods focus on exploring fine-grained image features and incorporate transformer-based complex fusion strategies, while ignoring the heavy computational burden. Recently, some lightweight multilayer perceptrons (MLP)-based methods have been successfully applied to multimodal sentiment classification tasks. In this article, we propose an effective rearrangement and restore mixer model (RR-Mixer) for TMSC, which dedicates the interaction of image, text, and targets along the modal-axis, sequential-axis, and feature channel-axis through rearrangement and restore operations. Specifically, we take vision transformer (ViT) and robustly optimized BERT (RoBERTa) pretrained models to extract image and textual features, respectively. Further, we adopt cosine similarity to select the most semantically relevant image features. Then, an RR-Mixer module is designed for mixed multimodal features, with the core technology consisting of rolling, grouping rearrangement and restore operations. Moreover, we introduce MLP Unit to learn the information of different modalities for intermodal interaction. The results show that our model achieves superior performance on two benchmark multimodal datasets, TWITTER-15 and TWITTER-17, with a significant improvement of 4.66% and 1.26% in terms of macro-F1. © 2020 IEEE.

关键词： Feature extraction

来源：评论

学校读者我要写书评

暂无评论

Target Detection Algorithm of SAR Ship Image Based on Improved Yolov8n 6

Target Detection Algorithm of SAR Ship Image Based on Improv...

引用

6th International Conference on Robotics and computer Vision, ICRCV 2024

作者： Lu, Qin Zhao, Tongzhou Hu, Wei Chai, Zhengpeng School of Computer Science and Engineering & Artificial Intelligence Wuhan Institute of Technology Wuhan China

ISBN: (纸本)9798331527426

With regard to ship detection in SAR images, there are problems that the ship scale changes greatly and the background environment is complex in SAR images, which leads to the low accuracy of ship detection. In order to solve the above problems, this paper proposes a target detection algorithm for SAR ship images based on the improved Yolov8n. Firstly, by introducing BiFPN (bi-directional feature pyramid network) into Yolov8n, cross-scale feature fusion is realized, and the model pays more attention to information of different scales. Secondly, LSK (Large Separable Kernel Attention) attention mechanism is added after the SPPF network of the model, which can capture more ship information and reduce the computational load in the complex background environment. Finally, Wise- IoU v2 is used to replace the original CIoU function in Yolov8n to improve the detection accuracy. The improved performance of the model is verified by using public data sets SAR-Ship-dataset dataset and SSDD dataset. The experimental results show that the improved Yolo v8n ship detection model is better than Yolov8n algorithm model to a great extent. On the SAR-Ship-dataset dataset, mAP0.5 improves by 0.423% and mAP0.5-0.95 improves by 2.362%;On the SSDD dataset, mAP0.5 improves by 0.551 % and mAP0.5-0.95 improves by 3.147%. © 2024 IEEE.

关键词： Radar imaging

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：