检索结果-内蒙古大学图书馆

An intelligent image detection method using improved canny edge detection operator

International Journal of Information technology and Management 2022年第4期21卷 369-381页

作者： Wang, Qian Chen, Wenxia Peng, Haiyun College of Computer Science and Technology Zhoukou Normal University Henan Zhoukou466001 China Zhoukou Senior Technical School Henan Zhoukou466001 China

In order to meet the requirement of edge detection of paper disease image in papermaking process, a paper disease image detection method based on improved Canny operator is proposed. Firstly, according to the principle of Gauss filtering and the method of feature statistical analysis, the filtering function and window are selected adaptively. Then, in the gradient solution, the traditional 2 × 2 neighbourhood is replaced by the 3 × 2 or 2 × 3 neighbourhood which enhances the weight of the intermediate pixel, and the accuracy of edge detection is improved by enhancing the influence of the intermediate pixel. Finally, the iterative averaging method is used to determine the optimal threshold and reduce the error rate of image edge segmentation. The experimental results show that this method can effectively detect the edges of paper disease area and has good edge continuity. Copyright © 2022 Inderscience Enterprises Ltd.

关键词： Statistical methods

来源：评论

学校读者我要写书评

暂无评论

Research on Multi-Model Fusion Method for Similar Multi-domain 5

Research on Multi-Model Fusion Method for Similar Multi-doma...

引用

5th International Conference on Robotics and computer Vision, ICRCV 2023

作者： Zhao, Junhua Song, Qingzeng Xue, Yongjiang Yu, Jing School of Computer Science and Technology Tiangong University Tianjin China School of Computer Science and Technology Tiangong University Haihe Lab of ITAI Tianjin China School of Telecommunication Engineering Beijing Polytechnic Beijing China

ISBN: (纸本)9798350326369

In order to improve the classification accuracy of test samples from different datasets in multi-domain image classification, this paper proposes a multi-domain sample classification method based on Baidu AI open platform image recognition and multi-model fusion. The classification task is divided into two steps: first, the domain to which the sample belongs is determined based on the predicted label and its probability output by the Baidu API, and then the deep learning model trained on that domain is used to predict the label of the sample. The effectiveness of the proposed method is verified through extensive experiments on three public datasets: Cifar-10, Cifar-100, and Mini-ImageNet. In addition, compared with existing algorithms, the proposed method shows significant improvements in classification performance. © 2023 IEEE.

关键词： Image classification

来源：评论

学校读者我要写书评

暂无评论

ChemGenX: AI in the Chemistry Classroom 24

ChemGenX: AI in the Chemistry Classroom

引用

2024 International Symposium on Artificial Intelligence for Education, ISAIE 2024

作者： Abbas, Touqeer Javed, Umair Mehmood, Faisal Raza, Muneeb Li, Hui Department of Computer Science and Technology Beijing University of Chemical Technology Beijing China School of Electrical and Information Engineering Zhengzhou University Henan Zhengzhou China Beijing China

ISBN: (纸本)9798400707100

Generative AI is reshaping education by offering personalized learning experiences and innovative teaching methods. In this study, we introduced ChemGenX, an AI tool designed to enhance chemistry education by generating molecular structures with user-defined features. The primary objective is to simplify molecular design and improve student learning. Our approach utilizes the coherently aligned multimodal semantic space using pre-trained CLIP;generate text features from image features to seamlessly alleviate the need for text conditioning, resulting in accurate predictions of bond lengths, angles, and molecular structures. The model was trained on a comprehensive molecular dataset, achieving competitive performance in zero-shot text-to-image conversion. Pilot studies with high school students and teachers demonstrate its potential in generating accurate molecular structures, including key features like bond lengths and angles. © 2024 Copyright held by the owner/author(s). Publication rights licensed to ACM.

关键词： Bond length

来源：评论

学校读者我要写书评

暂无评论

State Monitoring and Diagnosis of Autonomous Underwater Vehicles Based on Transformer 5

State Monitoring and Diagnosis of Autonomous Underwater Vehi...

引用

5th International Conference on System Reliability and Safety Engineering, SRSE 2023

作者： Wang, Jing Zhao, Xueqing Pei, Shicheng Luo, Xiaoyu Xi'an Polytechnic University School of Computer Science Technology Xi'an China University of Electronic Science and Technology of China Glasgow College Chengdu China

ISBN: (纸本)9798350305944

Autonomous underwater vehicles (AUVs) are extensively utilized in the examination of oceanic regions owing to their remarkable capabilities in long-range navigation and autonomous operations. Fault detection and diagnosis have always been important in maintaining AUV safety and reliability. In order to ensure that the AUVs can perform their tasks safely and effectively in a complex and unpredictable underwater environment, we propose a Transformer-based AUV fault diagnosis method (Trans-CNN), which can be used to accurately monitor and diagnose the status of the AUVs through sensor data. First, Trans-CNN adopts multi-layer convolutional neural networks (CNNs) to realize local information learning of multivariate sensor data. Subsequently, considering the limitations of CNN global feature learning, Trans-CNN further introduces the Transformer architecture, which realizes the global feature encoding of multi-directional information by adopting a multi-head attention mechanism. Therefore, the proposed Trans-CNN fully integrates the advantages of CNN and Transformer in global and local feature learning so as to comprehensively extract valuable fault feature information of autonomous underwater vehicles. Experiments on real AUVs datasets show that this method can accurately and effectively monitor and diagnose the fault type of AUVs and has better performance than existing fault diagnosis methods. This confirms that the fault diagnosis method has the promise of practical deep-sea applications. © 2023 IEEE.

关键词： Failure analysis

来源：评论

学校读者我要写书评

暂无评论

TELEX: Two-Level Learned Index for Rich Queries on Enclave-based Blockchain Systems

引用

IEEE Transactions on Knowledge and Data Engineering 2025年第7期37卷 4299-4313页

作者： Wu, Haotian Tang, Yuzhe Shen, Zhaoyan Tao, Jun Lin, Chenhao Peng, Zhe The Hong Kong Polytechnic University Department of Industrial and Systems Engineering Hong Kong Syracuse University Department of Electrical Engineering and Computer Science United States Shandong University School of Computer Science and Technology China Southeast University School of Cyber Science and Engineering China Xi'an Jiaotong University Faculty of Electronic and Information Engineering China The Hong Kong Polytechnic University Shenzhen Research Institute Hong Kong

Blockchain has become a popular paradigm for secure and immutable data storage. Despite its numerous applications across various fields, concerns regarding the user privacy and result integrity during data queries persist. Additionally, the need for rich query functionalities to harness the full potential of blockchain data remains an area ripe for exploration. In order to address these challenges, our paper first utilizes a framework based on the Trusted Execution Environment (TEE) and oblivious RAM technique to achieve both privacy and data integrity. To enhance the query efficiency over the entire blockchain, we then devise a two-level learned indexing methodology named TELEX within the TEE for both integer and string keys. We also propose different query processing algorithms for versatile query types, including exact queries, aggregate queries, Boolean queries, and range queries. By implementing the prototype and conducting extensive evaluation, we demonstrate the feasibility and remarkable improvement in efficiency compared to existing solutions. © 1989-2012 IEEE.

关键词： Query languages

来源：评论

学校读者我要写书评

暂无评论

Learning local line descriptors with fcnns and transfer learning by minimizing triplet margin loss

UPB Scientific Bulletin, Series C: Electrical Engineering an...

引用

UPB Scientific Bulletin, Series C: Electrical Engineering and computer science 2021年第2期83卷 133-144页

作者： Huo, Zhanqiang Fu, Miaomiao Luo, Fen Liu, Yujie School of Computer Science and Technique Henan Polytechnic University Jiaozuo454003 China

In this paper, it is demonstrated how to learn line feature descriptors for line matching from image data, which is very important for many image processing and computer vision applications. In order to encode these functions, the FCNNs (Fully Convolutional Neural Networks) based models are developed. The main contributions of this method can be summarized form four aspects: (1) it describes the line features with mean patches and standard deviation patches;this idea comes from MSLD (mean–standard deviation line descriptor);(2) three FCNNs based models are introduced, and the central-surrounding to data augmentation is used before training;(3) the parameters of the pre-trained L2-Net are transferred to the three models to solve the problem of insufficient amount of training data;(4) the triplet margin loss is used to optimize these three models. The proposed method is verified by the experiments on the Oxford dataset and other datasets, and the experimental results show that the line feature descriptors obtained by our three models have better robustness under the conditions of viewpoint change, rotation change, blur change, and scale change and many others. © 2021, Politechnica university of Bucharest. All rights reserved.

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

Asymptotics-Aware Multi-View Subspace Clustering

引用

IEEE Transactions on Multimedia 2025年 27卷 3650-3663页

作者： Xu, Yesong Chen, Shuo Li, Jun Yang, Jian Anhui Polytechnic University School of Computer and Information Wuhu241000 China RIKEN Center for Advanced Intelligence Project Tokyo103-0027 Japan Nanjing University of Science and Technology PCA Laboratory Key Laboratory of Intelligent Perception and Systems for High-Dimensional Information of Ministry of Education Nanjing210094 China Nanjing University of Science and Technology Jiangsu Key Laboratory of Image and Video Understanding for Social Security School of Computer Science and Engineering Nanjing210094 China

Recently, multi-view subspace clustering has attracted extensive attention due to the rapid increase of multi-view data in many real-world applications. The main goal of this task is to learn a common representation of multiple subspaces from the given multi-view data, and most existing methods usually directly merge multiple groups of features by the single-step integration. However, there may exist large disparities among different views of the data, and thus the conventional single-step practice can hardly obtain a generally consistent feature representation for the multi-view data. To overcome this challenge, we present a novel approach dubbed "Asymptotics-Aware Multi-view Subspace Clustering (A2 MSC)"to pursue a consistent feature representation in a multi-step way, which iteratively conducts the data recovery to gradually reduce the differences between pairwise views. Specifically, we construct an asymptotic learning rule to update the feature representation, and the iteration result converges to a consistent feature vector for characterizing each instance of the original multi-view data. After that, we utilize such a new feature representation to learn a clustering-oriented similarity matrix via minimizing a self-expressive objective, and we also design the corresponding optimization algorithm to solve it with convergence guarantees. Theoretically, we prove that the learned asymptotic representation effectively integrates multiple views, thereby ensuring the effective handling of multi-view data. Empirically, extensive experimental results demonstrate the superiority of our proposed A2 MSC over the state-of-the-art multi-view subspace clustering approaches. © 2025 IEEE.

关键词： Data assimilation

来源：评论

学校读者我要写书评

暂无评论

PrivShape: Extracting Shapes in Time Series under User-Level Local Differential Privacy 40

PrivShape: Extracting Shapes in Time Series under User-Level...

引用

40th IEEE International Conference on Data Engineering, ICDE 2024

作者： Mao, Yulian Ye, Qingqing Hu, Haibo Wang, Qi Huang, Kai Southern University of Science and Technology Department of Computer Science and Engineering China The Hong Kong Polytechnic University Department of Electrical and Electronic Engineering Hong Kong Research Institute of Trustworthy Autonomous Systems Southern University of Science and Technology China School of Computer Science and Engineering Macau University of Science and Technology China

ISBN: (纸本)9798350317152

Time series have numerous applications in finance, healthcare, IoT, and smart city. In many of these applications, time series typically contain personal data, so privacy infringement may occur if they are released directly to the public. Recently, local differential privacy (LDP) has emerged as the state-of-the-art approach to protecting data privacy. However, existing works on LDP-based collections cannot preserve the shape of time series. A recent work, PatternLDP, attempts to address this problem, but it can only protect a finite group of elements in a time series due to ?-event level privacy guarantee. In this paper, we propose PrivShape, a trie-based mechanism under user-level LDP to protect all elements. PrivShape first transforms a time series to reduce its length, and then adopts trie-expansion and two-level refinement to improve utility. By extensive experiments on real-world datasets, we demonstrate that PrivShape outperforms PatternLDP when adapted for offline use, and can effectively extract frequent shapes. © 2024 IEEE.

关键词： Time series

来源：评论

学校读者我要写书评

暂无评论

SmartNet: One-shot Talking Head Synthesis via Subtle Motion and Appearance Compensation

SmartNet: One-shot Talking Head Synthesis via Subtle Motion ...

引用

2025 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2025

作者： Hu, Wei Ji, Yuzhu Zeng, An Pan, Dan Zhang, Yiqun Zhang, Haijun School of Computer Science and Technology Guangdong University of Technology Guangzhou China Guangdong Polytechnic Normal University Guangzhou China Harbin Institute of Technology Shenzhen China

ISBN: (纸本)9798350368741

One-shot talking head synthesis aims to animate a source person's portrait with driving video sequences. Recent facial keypoint-based methods have achieved remarkable animation performance and produced high-quality results. However, it remains challenging to perform cross-identity face reenactment by transferring subtle facial motions with correct geometry and appearance. To break the above limitations, in this paper, we propose a subtle motion compensation network to recover correct facial expressions by leveraging the decoupled 3D Morphable Model (3DMM) coefficient. In addition, to generate faithful animation results, a facial appearance feature memory bank is designed to learn accurate facial features and better recover the appearance. Experimental results have demonstrated that our proposed model can outperform state-of-the-art methods by generating faithful videos with correct subtle motion transfer and consistent identity preserving. © 2025 IEEE.

关键词： Face animation Feature compensation Motion transfer Talking head synthesis

来源：评论

学校读者我要写书评

暂无评论

DPLink: Link Prediction Method Based on Local Structure of Drug-Protein Network

DPLink: Link Prediction Method Based on Local Structure of D...

引用

International Conference on Wireless Communications, Networking and Applications (WCNA 2022)

作者： Zhang, Bofeng Li, Haiyan Lv, Hehe School of Computer Science and Technology Kashi University Kashi844006 China School of Computer and Information Engineering Shanghai Polytechnic University Shanghai201209 China

ISBN: (纸本)9789819939503

In silico methods for predicting drug-protein interactions (DPI) play an important role in the process of new discovery of drug actions. A popular method for DPI prediction is based on network structural similarity, which requires only drug and protein interaction relationships. However, the prediction method based on structural similarity cannot accurately predict the DPI in some network structures, for example, there are differences in local structures, but the similarity prediction method calculates the same result. A new DPI prediction method (DPLink) is proposed, which utilizes the local structural information of direct interactions to calculate the likelihood of interactions between drug and protein. The integration results of the existing method and DPLink show that the method is capable of achieving more accurate performance than structure-based prediction methods. © 2023, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.

关键词： Proteins

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：