检索结果-内蒙古大学图书馆

23rd ieee/CVF Winter Conference on Applications of Computer Vision (WACV)

作者： Chen, Wei-Chi Chu, Wei-Ta Natl Cheng Kung Univ Tainan Taiwan

ISBN: (纸本)9781665493468

With labeled data, self distillation (SD) has been proposed to develop compact but effective models without a complex teacher model available in advance. Such approaches need labeled data to guide the self distillation process. Inspired by self-supervised (SS) learning, we propose a self-supervised self distillation (SSSD) approach in this work. Based on an unlabeled image dataset, a model is constructed to learn visual representations in a self-supervised manner. This pre-trained model is then adopted to extract visual representations of the target dataset and generates pseudo labels via clustering. The pseudo labels guide the SD process, and thus enable SD to proceed in an unsupervised way (no data labels are required at all). We verify this idea based on evaluations on the CIFAR-10, CIFAR-100, and ImageNet-1K datasets, and demonstrate the effectiveness of this unsupervised SD approach. Performance out-performing similar frameworks is also shown.

关键词： visualization Computer vision Computational modeling Clustering algorithms Self-supervised learning Feature extraction data models

来源：评论

学校读者我要写书评

暂无评论

Supervised Detail-Guided Multiscale State-Space Model for Pan-Sharpening

引用

ieee TRANSACTIONS ON GEOscience AND REMOTE SENSING 2025年 63卷

作者： Zhao, Mengjiao Ma, Mengting Li, Xiangdong Ma, Xiaowen Zhang, Wei Song, Siyang Zhejiang Univ Sch Software Technol Hangzhou 310027 Peoples R China Zhejiang Univ Sch Comp Sci & Technol Hangzhou 310027 Peoples R China Zhejiang Univ Innovat Ctr Yangtze River Delta Jiaxing 314103 Zhejiang Peoples R China Univ Exeter Dept Comp Sci Exeter EX4 4PY England

Pan-sharpening reconstructs the high-resolution multispectral (HR-MS) image from its corresponding panchromatic (PAN) image and low-resolution multispectral (LR-MS) image. However, existing deep learning (DL)-based pan-sharpening methods typically suffer from three challenges: 1) the vanilla LR-MS image upsampling employed by them fails to consider domain knowledge, thereby disregarding crucial information;2) while remote sensing images exhibit multiscale complex land features, existing methods fail to fully exploit crucial multiscale spatial information, that is, scale transformation layers in their models are not effective;and 3) existing convolutional neural network (CNN) and transformer-based pan-sharpening backbones are constrained by inherent local receptive fields or quadratic computational complexity, making them difficult to balance their effectiveness and efficiency. To address these issues, we propose a novel supervised detail-guided multiscale state-space model for pan-sharpening, namely SDMSPan. Our SDMSPan consists of three residual state-space modules (Res-SSMs) that are responsible for handling image information at three spatial scales, where each Res-SSM aims to model both local and long-range dependencies between PAN and LR-MS images at a specific spatial scale with lower computational cost. Between each pair of Res-SSMs, a novel detail-guided upsampling block (DGUB) is proposed to apply spatial details of the PAN image to guide effective and task-aware intermediate feature upsampling, where a novel multiscale intermediate spatial-spectral supervision strategy is also proposed to supervise the training of every DGUB. Experimental results demonstrate that our proposed approach significantly outperforms other state-of-the-art methods in performance. Our code is provided at https://***/zhaomengjiao123/SDMSPan.

关键词： Feature extraction Computational modeling Distortion Transformers Spatial resolution Computational efficiency Optimization data mining Remote sensing visualization Pan-sharpening remote sensing state-space model

来源：评论

学校读者我要写书评

暂无评论

Learning Pose Controllable Human Reconstruction With Dynamic Implicit Fields From a Single Image

引用

ieee TRANSACTIONS ON visualization AND COMPUTER GRAPHICS 2025年第2期31卷 1389-1401页

作者： Li, Jituo Liu, Xinqi Lu, Guodong Zhejiang Univ Inst Design Engn Sch Mech Engn Hangzhou 310027 Peoples R China Zhejiang Univ Robot Inst Hangzhou 310027 Peoples R China

Recovering a user-special and controllable human model from a single RGB image is a nontrivial challenge. Existing methods usually generate static results with an image consistent subject's pose. Our work aspires to achieve pose-controllable human reconstruction from a single image by learning a dynamic (multi-pose) implicit field. We first construct a feature-embedded human model (FEHM) as a bridge to propagate image features to different pose spaces. Based on FEHM, we then encode three pose-decoupled features. Global image features represent user-specific shapes in images and replace widely used pixel-aligned ways to avoid unwanted shape-pose entanglement. Spatial color features propagate FEHM-embedded image cues into 3D pose space to provide spatial high-frequency guidance. Spatial geometry features improve reconstruction robustness by using the surface shape of the FEHM as the prior. Finally, new implicit functions are designed to predict the dynamic human implicit fields. For effective supervision, a realistic human avatar dataset, SimuSCAN, with 1000+ models is constructed using a low-cost hierarchical mesh registration method. Extensive experiments demonstrate that our method achieves the state-of-the-art reconstruction level.

关键词： Image reconstruction Biological system modeling Shape Image color analysis Task analysis Avatars Clothing Clothed human reconstruction dynamic implicit fields neural implicit representation single image human avatar dataset

来源：评论

学校读者我要写书评

暂无评论

GeoBeeDashboard: An Intelligent Geospatial Platform for Enhancing Waste Management in Tangerang City 13

GeoBeeDashboard: An Intelligent Geospatial Platform for Enha...

引用

13th ieee International Conference on Communication, Networks and Satellite, COMNETSAT 2024

作者： Muljana, Richie Deffansyah, Syauqi Akmal Andrew, Hari Gunawan, Alexander Agung Santoso Irwansyah, Edy Pramudya, Fabian Surya Ario, Muhamad Keenan Syahputra, Muhammad Edo Achmad, Said School of Computer Science Bina Nusantara University Data Science Program Computer Science Department Jakarta11480 Indonesia Bina Nusantara University School of Computer Science Computer Science Department Jakarta11480 Indonesia Bina Nusantara University School of Computer Science Mathematics Department Jakarta11480 Indonesia

ISBN: (纸本)9798350368086

In this study, we created an Intelligent Geospatial Platform called GeoBeeDashboard. The GeoBeeDashboard combines advanced geospatial technologies and artificial intelligence (AI), allowing real-time data visualization and optimized decision-making across multiple sectors. The system architecture consists of a ***-based front end, a FastAPI-driven back end, and PostGIS for geospatial data management. One of the GeoBeeDashboard's primary use cases is addressing waste management issues in Tangerang City, Indonesia. As rapid urbanization increases waste generation, effective management becomes critical. The dashboard identifies critical trends and inefficiencies in Tangerang's districts by K-Means clustering waste management data from 2020 to 2023. Our findings identify distinct clusters, highlighting areas with effective waste management and those that require targeted interventions. The benefits that local government can get from GeoBeeDashboard by utilising insights based on clustering analysis of data per each district in Tangerang to optimise waste collection, allocate resources efficiently, and improve the policy-making process for each district in Tangerang. The GeoBeeDashboard improves local government capabilities by providing a scalable model for other cities facing similar urban challenges. In conclusion, GeoBeeDashboard provides local governments with a powerful tool for addressing waste management inefficiencies and fostering sustainable urban development. © 2024 ieee.

关键词： Urban planning

来源：评论

学校读者我要写书评

暂无评论

Design and Implementation of a Quality Big data Analysis Platform for Multi-Structure data Throughout the Entire Lifecycle of Integrated Engines 2

Design and Implementation of a Quality Big Data Analysis Pla...

引用

2nd International Conference on Machine Learning, Cloud Computing, and Intelligent Mining, MLCCIM 2023

作者： Yang, Shuaihu Feng, Min Xiao, Min Computer and Electronic Information Science College Guangxi University Nanning China Guangxi Yuchai Power Corporation Information Technology Department Yulin China

ISBN: (纸本)9798350328530

With the development and popularization of computer technology, computers have played an important role in improving production efficiency. In the face of opportunities and challenges in the new era, China has established the goal of transitioning from 'manufacturing' to 'intelligent manufacturing', integrating intelligent technology into production, improving production efficiency, and gaining an advantage in international competition. In this context, Yuchai, as a leading enterprise in the engine manufacturing industry, has the task of developing intelligent manufacturing and improving enterprise competitiveness. Therefore, this article takes Yuchai's multi-structure data throughout the entire lifecycle of the engine as the source data and designs and implements a quality big data analysis platform. By integrating data from production, inspection, sales, operation, and after-sales, big data analysis of the engine is achieved and the analysis results are visualized, thereby providing guidance for production and improving product quality. Meanwhile, a similarity clustering-based engine performance prediction algorithm is proposed. The system can display the distribution of failures in different regions through data visualization solutions such as heat maps, and achieve better allocation of components by adjusting engine component parameters before leaving the factory according to local natural environment and working conditions. The application and implementation of the system are the result of the integration and application of digital assets, and have extensive application value in engine production and the aftermarket, directly and indirectly bringing huge economic benefits. © 2023 ieee.

关键词： Clustering algorithms

来源：评论

学校读者我要写书评

暂无评论

FMRT: Learning Accurate Feature Matching With Reconciliatory Transformer

引用

ieee TRANSACTIONS ON AUTOMATION science AND ENGINEERING 2025年 22卷 11826-11842页

作者： Wang, Li Zhang, Xinyu Jiang, Zhiqiang Dai, Kun Xie, Tao Yang, Lei Yu, Wenhao Shen, Yang Xu, Bin Li, Jun Beijing Inst Technol Sch Mech Engn Beijing 100081 Peoples R China Tsinghua Univ Sch Vehicle & Mobil Beijing 100084 Peoples R China Tsinghua Univ Suzhou Automot Res Inst Suzhou 215200 Peoples R China Natl Univ Singapore Dept Elect & Comp Engn Singapore 119077 Singapore Harbin Inst Technol State Key Lab Robot & Syst Harbin 150006 Peoples R China Tsinghua Univ State Key Lab Automot Safety & Energy Beijing 100084 Peoples R China

Local Feature Matching, a pivotal component of numerous computer vision tasks (e.g., structure from motion and visual localization), has been effectively addressed by Transformer-based methods. Nevertheless, these methods solely incorporate long-range context information among keypoints with a fixed receptive field, which constrains the network from appropriately reconciling the importance of features with diverse receptive fields to realize complete image perception, hence limiting feature matching accuracy. In addition, these methods employ a conventional handcrafted encoding approach to incorporate positional information of keypoints into visual descriptors, which limits the capability of networks to extract effective positional encoding message. In this study, we propose FMRT, a novel detector-free method that reconciles local features with diverse receptive fields adaptively and utilizes parallel networks to realize reliable positional encoding. Specifically, FMRT proposes a dedicated reconciliatory transformer (RecFormer) that contains a global perception attention layer to identify visual descriptors with different receptive fields and integrate global context information under various scales, a perception weight layer to measure the importance of various receptive fields adaptively, and a local perception feed-forward network to extract deep aggregated multi-scale local feature representation. Moreover, we introduce a novel axis-wise position encoder (AWPE) that views positional encoding as two keypoints encoding tasks along the row and column dimensions, decouples the x- and y-coordinates of keypoints into two independent 1D vectors, and designs two parallel network branches to explicitly encodes geometric correlations among keypoints, hence realizing reliable positional encoding. Extensive experiments indicate that FMRT yields impressive performance on multiple tasks, including relative pose estimation, visual localization, homography estimation, and image

关键词： Feature extraction visualization Transformers Encoding data mining Location awareness Correlation Computational modeling Accuracy Vectors Local feature matching transformer deep learning position encoder

来源：评论

学校读者我要写书评

暂无评论

Implementation of Double Diamond Design Model for User Interface of Bandung City Dashboard 8

Implementation of Double Diamond Design Model for User Inter...

引用

8th International Conference on Information Technology and Digital Applications, ICITDA 2023

作者： Hanifah, Aliffia Aditya, Bayu Rima Mulyawan, Muhammad School of Applied Science Telkom University Bandung Indonesia Diskominfo Kota Bandung Bandung Indonesia

ISBN: (纸本)9798350344691

The Bandung City Dashboard website is one of the ICT services owned by the Bandung City Communication and Information Office as media providing data visualization. However, it was found that this website was not user-friendly, not attractive, and required adding new content. Therefore, to increase users' engagement with the website, the study aims to evaluate and improve the interface design by using the double diamond method including discover, define, develop, and deliver. The solution provided is to involve users in the website evaluation process and improve the interface design. The result of this study is an improved interface design based on measures of user preference before and after with three success indicators: effectiveness (before 83 %, after 100%), efficiency (before 69%, after 100%), and satisfaction (before 64%, after 85%). Thus, it can be concluded that the new interface design is better and more acceptable in referencing to user perception, experience, and usability. © 2023 ieee.

关键词： User interfaces

来源：评论

学校读者我要写书评

暂无评论

Designing Intelligent Intrusion Detection System Using Deep Learning 14

Designing Intelligent Intrusion Detection System Using Deep ...

引用

14th International Conference on Computing Communication and Networking Technologies, ICCCNT 2023

作者： Wasnik, Priyanka Chavhan, Nekita G H Raisoni College of Engineering Department of Computer Science and Engineering Nagpur India G H Raisoni College of Engineering Department of Data Science IoT and Cyber Security Nagpur India

ISBN: (纸本)9798350335095

In order to increase the effectiveness and precision of cyber threat detection in computer networks, deep learning techniques are being applied in the construction of intelligent intrusion detection systems (IDS). The suggested IDS can automatically learn and recognize complicated patterns and characteristics from network traffic data by utilizing the capabilities of deep neural networks. Convolutional neural networks (CNNs), recurrent neural networks (RNNs), or deep belief networks (DBNs) are some examples of deep learning architectures that can be built as part of the system design. These models can then be trained on labelled data to distinguish between legitimate and nefarious network behavior. In order to detect both well-known and new threats, the IDS concentrates on capturing both known attack signatures and unusual behaviors. Deep learning models are well suited for real-time intrusion detection in dynamic network environments because they can generalize and adapt to changing threats. By using methods like transfer learning, dimensionality reduction, and visualization, the system also handles issues like feature extraction, scalability, and interpretability. The suggested IDS seeks to achieve higher detection rates, lower false positive rates, and improved resilience against evasion strategies through thorough experimentation and review. The ultimate objective is to create an intelligent IDS that accurately detects and mitigates hostile activity to successfully protect computer networks, improving the overall security posture of the network infrastructure. © 2023 ieee.

关键词： Intrusion detection

来源：评论

学校读者我要写书评

暂无评论

Toward Micro Lip Reading in the Wild: dataset, Theory, and Practices

引用

ieee TRANSACTIONS ON AUTOMATION science AND ENGINEERING 2025年 22卷 11794-11807页

作者： Wang, Ran Cheng, Yingsong Cao, Anqi Zhong, Shengtao Liao, Xiang He, Yanjiu Ye, Han Sun, Chutian Zhao, Qinlao Du, Min Xiao, Yang Huazhong Univ Sci & Technol Sch Journalism & Informat Commun Wuhan 430074 Peoples R China Huazhong Univ Sci & Technol Sch Future Technol Wuhan 430074 Peoples R China Wuhan Univ Sch Elect Informat Wuhan 430072 Peoples R China Huazhong Univ Sci & Technol Sch Comp Sci & Technol Wuhan 430074 Peoples R China Univ Calif Irvine Donald Bren Sch Informat & Comp Sci Irvine CA 92697 USA Peking Univ Sch Software & Microelect Beijing 100871 Peoples R China Huazhong Univ Sci & Technol Sch Artificial Intelligence & Automat Natl Key Lab Multispectral Informat Intelligent Pr Wuhan 430074 Peoples R China Huazhong Univ Sci & Technol Sch Mech Sci & Engn Wuhan 430074 Peoples R China ByteDance Beijing 100089 Peoples R China

Micro lip reading, is characterized by tiny lip movements when someone is speaking. It is highly preferred in robotics and automation, such as a social robot for geriatric care, a patrol robot system for hospitals, and speech system for hearing impaired individuals, etc. However, it has not been well studied before. In this paper, we shed the light to this research topic. A labelled micro lip reading dataset (i.e. HUST-LMLR) of 399 video samples is first established. The samples are captured from the unconstrained movies. One key challenge for micro lip reading in the wild lies on extracting fine features of lip movements effectively and robustly. In this paper, we pay the research efforts to address this issue from two aspects: facial context attention and feature extraction. First, we propose a multi-task learning model for micro lip reading in the wild for the first time. It concerns micro lip reading and facial landmark detection jointly for capturing global face context with soft attention, to facilitate micro lip reading. Second, we propose that motion feature should be concerned and combined with appearance feature jointly to characterize tiny lip movements effectively. Finally, the experiments on HUST-LMLR demonstrate the challenges of our dataset and our proposed approach results in a remarkable improvement of the state-of-the-art by over 26% WER in HUST-LMLR. Furthermore, results in a well-known public dataset LRS2 also show the generalization and superiority of our approach. If accepted, we will publish our HUST-LMLR dataset and relative supporting matierials at https://***/Micro-lip-reading/.

关键词： Lips Feature extraction Automation visualization Faces data mining Multitasking Motion pictures Location awareness Electronic mail Micro lip reading spatial-temporal pattern recognition appearance and motion deep learning

来源：评论

学校读者我要写书评

暂无评论

Revolutionizing Agriculture: An IoT-Driven ML-Blockchain Framework 5.0 for Optimal Crop Prediction 5

Revolutionizing Agriculture: An IoT-Driven ML-Blockchain Fra...

引用

5th International Conference on Sustainable Technologies for Industry 5.0, STI 2023

作者： Sizan, Najmus Sakib Dey, Diganta Mia, Md. Solaiman Layek, Md. Abu Jagannath University Department of Computer Science and Engineering Bangladesh Green University of Bangladesh Department of Computer Science and Engineering Bangladesh

ISBN: (纸本)9798350394290

With cutting-edge technology development, the agricultural landscape is experiencing a revolutionary transformation. The Internet of Things (IoT)-driven Machine Learning (ML)- Blockchain framework 5.0 for agriculture presented in this paper is a ground-breaking system created to revolutionize crop forecasting and provide farmers with data-driven insights. The framework continually collects realtime data from sensor networks, utilizing the capabilities of the IoT to monitor critical environmental indicators and soil nutrient levels. ML techniques are used to analyze this extensive information, resulting in precise crop projections and customized insights for effective resource management. In this paper, we have experienced that the Support Vector Machine (SVM) model's prediction got an amazing accuracy for crops which is 97.73%. Blockchain technology plays a critical role in this system, providing a decentralized, tamper-proof ledger that ensures data transparency and integrity, bolstering credibility and facilitating collective decision-making among stakeholders, thereby cultivating trust and accountability within the agricultural community. Moreover, to enhance accessibility and engagement, a user-friendly online application has been created to enable stakeholders to access and examine crop projections and historical sensor data, facilitating user engagement and data visualization. The suggested framework imagines a new age in agriculture, where sustainable farming methods and precision farming boost crop yields and give agricultural communities more authority. © 2023 ieee.

关键词： Internet of things

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：