检索结果-内蒙古大学图书馆

Supervised deep learning for content-aware image retargeting with Fourier Convolutions

Multimedia Tools and Applications 2024年第36期83卷 83611-83627页

作者： Givkashi, MohammadHossein Naderi, MohammadReza Karimi, Nader Shirani, Shahram Samavi, Shadrokh Department of Electrical and Computer Engineering Isfahan University of Technology Isfahan84156-83111 Iran Department of Electrical and Computer Engineering McMaster University HamiltonONL8S 4L8 Canada Computer Science Department Seattle University SeattleWA98122 United States

Image retargeting aims to alter the size of the image with attention to the contents. One of the main obstacles to training deep learning models for image retargeting is the need for a vast labeled dataset. Labeled datasets are unavailable for training deep learning models in the image retargeting tasks. As a result, we present a new supervised approach for training deep learning models. We use the original images as ground truth and create inputs for the model by resizing and cropping the original images. A second challenge is generating different image sizes in inference time. However, normal convolutional neural networks cannot generate images of different sizes than the input image. To address this issue, we introduced a new method for supervised learning. In our approach, a mask is generated to show the desired size and location of the object. Then the mask and the input image are fed to the network. Comparing image retargeting methods and our proposed method demonstrates the model’s ability to produce high-quality retargeted images. Afterward, we compute the image quality assessment score for each output image based on different techniques and illustrate the effectiveness of our approach. © The Author(s), under exclusive licence to Springer science+Business Media, LLC, part of Springer Nature 2024.

关键词： Convolution

来源：评论

学校读者我要写书评

暂无评论

Swapped logit distillation via bi-level teacher alignment

引用

Multimedia Systems 2025年第3期31卷 1-13页

作者： Limantoro, Stephen Ekaputra Lin, Jhe-Hao Wang, Chih-Yu Tsai, Yi-Lung Shuai, Hong-Han Huang, Ching-Chun Cheng, Wen-Huang Department of Electrical Engineering and Computer Science National Yang Ming Chiao Tung University Hsinchu300093 Taiwan Cyberlink New Taipei City231 Taiwan Department of Computer Science and Information Engineering National Taiwan University Taipei10617 Taiwan

Knowledge distillation (KD) compresses the network capacity by transferring knowledge from a large (teacher) network to a smaller one (student). It has been mainstream that the teacher directly transfers knowledge to the student with its original distribution, which can possibly lead to incorrect predictions. In this article, we propose a logit-based distillation via swapped logit processing, namely Swapped Logit Distillation (SLD). SLD is proposed under two assumptions: (1) the wrong prediction occurs when the prediction label confidence is not the maximum;(2) the "natural" limit of probability remains uncertain as the best value addition to the target cannot be determined. To address these issues, we propose a swapped logit processing scheme. Through this approach, we find that the swap method can be effectively extended to teacher and student outputs, transforming into two teachers. We further introduce loss scheduling to boost the performance of two teachers’ alignment. Extensive experiments on image classification tasks demonstrate that SLD consistently performs best among previous state-of-the-art methods. © The Author(s), under exclusive licence to Springer-Verlag GmbH Germany, part of Springer Nature 2025.

关键词： Students

来源：评论

学校读者我要写书评

暂无评论

Explainable Depression Detection Using Large Language Models on Social Media Data 9

Explainable Depression Detection Using Large Language Models...

引用

9th Workshop on Computational Linguistics and Clinical Psychology, CLPsych 2024

作者： Wang, Yuxi Inkpen, Diana Buddhitha, Prasadith School of Electrical Engineering and Computer Science University of Ottawa Canada

ISBN: (纸本)9798891760806

Due to the rapid growth of user interaction on different social media platforms, publicly available social media data has increased substantially. The sheer amount of data and level of personal information being shared on such platforms has made analyzing textual information to predict mental disorders such as depression a reliable preliminary step when it comes to psychometrics. In this study, we first proposed a system to search for texts that are related to depression symptoms from the Beck’s Depression Inventory (BDI) questionnaire, and provide a ranking for further investigation in a second step. Then, in this second step, we address the even more challenging task of automatic depression level detection, using writings and voluntary answers provided by users on Reddit. Several Large Language Models (LLMs) were applied in experiments. Our proposed system based on LLMs can generate both predictions and explanations for each question. By combining two LLMs for different questions, we achieved better performance on three of four metrics compared to the state-of-the-art and remained competitive on the one remaining metric. In addition, our system is explainable on two levels: first, knowing the answers to the BDI questions provides clues about the possible symptoms that could lead to a clinical diagnosis of depression;second, our system can explain the predicted answer for each question. ©2024 Association for Computational Linguistics.

关键词： Social networking (online)

来源：评论

学校读者我要写书评

暂无评论

Dielectric Temporal Dynamics and Challenges Towards Accurate Skin Lesion Detection 6

Dielectric Temporal Dynamics and Challenges Towards Accurate...

引用

6th Australian Microwave Symposium, AMS 2025

作者： Raza Naqvi, Syed Akbar Abbosh, Amin The University of Queensland School of Electrical Engineering and Computer Science Queensland Australia

ISBN: (纸本)9798350379969

This paper presents the current challenges of expanding our handheld skin cancer detection system towards large-scale application on different skin regions across the human body. The analysis is based on temporal dielectric trend assessment of common moles and adjacent skin regions. The study indicates that developing a robust skin cancer detection system requires consideration of additional factors, such as hair strand interference compensation and dermal hydration levels in parallel with dielectric data acquisition. These biomechanical descriptor variations, including trans-epidermal water loss, contribute to fluctuations in the dielectric properties of the same skin region measured at different times. © 2025 IEEE.

关键词： Lung cancer

来源：评论

学校读者我要写书评

暂无评论

Poster: Platooning Revisited: What is the Personal Benefit Compared to ACC and Human Driving? 15

Poster: Platooning Revisited: What is the Personal Benefit C...

引用

15th IEEE Vehicular Networking Conference, VNC 2024

作者： Heinovski, Julian Ergenc, Doganalp Dressler, Falko School of Electrical Engineering and Computer Science TU Berlin Germany

ISBN: (纸本)9798350362701

Platooning has been researched for decades but debate about its lasting impact is still ongoing. Meanwhile, adaptive cruise control (ACC) became de facto standard for all new cars as well as for automated driving on the freeway. An evaluation of the personal benefit these systems offer remains difficult. To this end, we propose to assess driving systems by looking at the overall trip cost to incentivize drivers accordingly. For this, we define a novel metric to quantify the total trip cost, combining fuel cost and travel time within a monetary unit. We show the application of our new metric in a case study, comparing human driving, ACC, and platooning. Our results indicate that human driving always loses against ACC and that platooning has a significant advantage in mid to high traffic densities. © 2024 IEEE.

关键词： Adaptive cruise control

来源：评论

学校读者我要写书评

暂无评论

AR-Light: Enabling Fast and Lightweight Multi-User Augmented Reality via Semantic Segmentation and Collaborative View Synchronization

引用

IEEE Transactions on computers 2025年第6期74卷 2073-2086页

作者： Wen, Yu Shah, Aamir Bader Cao, Ruizhi Zhang, Chen Chen, Jiefu Wu, Xuqing Xie, Chenhao Fu, Xin University of Houston Department of Electrical and Computer Engineering HoustonTX77204 United States Beihang University School of Computer Science and Engineering Beijing100191 China University of Houston Department of Information Science Technology HoustonTX77204 United States

Multi-user Augmented Reality (MuAR) allows multiple users to interact with shared virtual objects, facilitated by exchanging environment information. Current MuAR systems rely on 3D point clouds for real-world analysis, view synchronization, object rendering, and movement tracking. However, the complexity of 3D point clouds leads to significant processing delays, with approximately 80% of overhead in commercial frameworks. This hampers usability and degrades user experience. Our analysis reveals that maintaining the facing side of the real-world scene in a stable environment provides sufficient information for virtual object placement and rendering. To address this, we introduce a lightweight quadtree structure, representing 2D scenes through semantic segmentation and geometry, as an alternative to 3D point clouds. Additionally, we propose a novel correction method to handle potential shifts in virtual object placement during view synchronization among users. Combining all designs, we implement a fast and lightweight MuAR framework named AR-Light and test our framework on commercial AR devices. The evaluation results on real-world applications demonstrate that AR-Light can achieve high performance in various real-world scenes while maintaining a comparable virtual object placement accuracy. © 1968-2012 IEEE.

关键词： Semantic Segmentation

来源：评论

学校读者我要写书评

暂无评论

Spatio-temporal Graph-Based Generation and Detection of Adversarial False Data Injection Evasion Attacks in Smart Grids

IEEE Transactions on Artificial Intelligence

引用

IEEE Transactions on Artificial Intelligence 2024年第12期5卷 6601-6616页

作者： Takiddin, Abdulrahman Ismail, Muhammad Atat, Rachad Serpedin, Erchin Florida State University Department of Electrical and Computer Engineering FAMU-FSU College of Engineering TallahasseeFL32310 United States Tennessee Tech University Department of Computer Science CookevilleTN38505 United States Lebanese American University Department of Computer Science and Mathematics Beirut1102-2801 Lebanon Texas A&M University Department of Electrical and Computer Engineering College StationTX77843 United States

Smart power grids are vulnerable to security threats due to their cyber-physical nature. Existing data-driven detectors aim to address simple traditional false data injection attacks (FDIAs). However, adversarial false data injection evasion attacks (FDIEAs) present a more serious threat as adversaries, with different levels of knowledge about the system, inject adversarial samples to circumvent the grid's attack detection system. The robustness of state-of-the-art graph-based detectors has not been investigated against sophisticated FDIEAs. Hence, this article answers three research questions. 1) What is the impact of utilizing spatio-temporal features to craft adversarial samples and how to select attack nodes? 2) How can adversaries generate surrogate spatio-temporal data when they lack knowledge about the system topology? 3) What are the required model characteristics for a robust detection against adversarial FDIEAs? To answer the questions, we examine the robustness of several detectors against five attack cases and conclude the following: 1) Attack generation with full knowledge using spatio-temporal features leads to 5%-26% and 2%-5% higher degradation in detection rate (DR) compared to traditional FDIAs and using temporal features, respectively, whereas centrality analysis-based attack node selection leads to 3%-11% higher degradation in DR compared to a random selection;2) Stochastic geometry-based graph generation to create surrogate adversarial topologies and samples leads to 3%-13% higher degradation in DR compared to traditional FDIAs;and 3) Adopting an unsupervised spatio-temporal graph autoencoder (STGAE)-based detector enhances the DR by 5-53% compared to benchmark detectors against FDIEAs. © 2020 IEEE.

关键词： Cyber attacks

来源：评论

学校读者我要写书评

暂无评论

Content-Aware Entity Alignment: Utilizing Structural and Semantic Similarities for Enhanced Inter-Knowledge Graph Integration 4

Content-Aware Entity Alignment: Utilizing Structural and Sem...

引用

4th International Conference on Digital Futures and Transformative Technologies, ICoDT2 2024

作者： Mehboob, Faiqa Satti, Fahad Ahmed Ali, Syed Imran Fraz, Muhammad Moazam NUST School of Electrical Engineering and Computer Science Islamabad Pakistan

ISBN: (纸本)9798350391862

Entity alignment in knowledge graphs is critical for maximizing the utility of interconnected data, especially in domains where data interoperability is essential. Entity alignment is the process of identifying and matching equivalent entities across different knowledge graphs. Traditional entity alignment methods often treat nodes as monolithic entities, which can lead to issues such as loss of granularity, semantic heterogeneity, and inadequate semantic understanding. These challenges reduce the effectiveness of the alignment process. Our research addresses these limitations by recognizing the heterogeneous nature of node information. We propose a novel approach to entity alignment that integrates structural similarities between nodes with their semantic and content information. This methodology aims to improve precision and context awareness in entity alignment, with potential applications in various fields, including healthcare, where efficient data exchange and collaborative research are crucial. © 2024 IEEE.

关键词： Knowledge graph

来源：评论

学校读者我要写书评

暂无评论

A New Multi Chaos-Based Compression Sensing Image Encryption

引用

computers, Materials & Continua 2023年第7期76卷 437-453页

作者： Fadia Ali Khan Jameel Ahmed Suliman A.Alsuhibany Department of Electrical Engineering Riphah International UniversityIslamabad44000Pakistan Department of Computer Science College of ComputerQassim UniversityBuraydah51452Saudi Arabia

The advancements in technology have substantially grown the size of image *** image encryption algorithms have limited capabilities to deal with the emerging challenges in big data,including compression and noise *** image encryption method that is based on chaotic maps and orthogonal matrix is proposed in this *** proposed scheme is built on the intriguing characteristics of an orthogonal *** Schmidt disperses the values of pixels in a plaintext image by generating a random orthogonal matrix using logistic chaotic *** the diffusion process,a block-wise random permutation of the data is performed using *** proposed scheme provides sufficient security and resilience to JPEG compression and channel noise through a series of experiments and security *** enables Partial Encryption(PE)for faster processing as well as complete encryption for increased *** higher values of the number of pixels change rates and unified average change intensity confirm the security of the encryption *** contrast to other schemes,the proposed approach can perform full and partial encryption depending on security requirements.

关键词： Chaos compression image encryption

来源：评论

学校读者我要写书评

暂无评论

Efficient Algorithms for Attributed Graph Alignment With Vanishing Edge Correlation

引用

IEEE Transactions on Information Theory 2025年第6期71卷 4556-4580页

作者： Wang, Ziao Wang, Weina Wang, Lele The University of British Columbia Department of Electrical and Computer Engineering VancouverBCV6T 1Z4 Canada University of Michigan Department of Electrical and Computer Engineering Ann ArborMI48109 United States Carnegie Mellon University Computer Science Department PittsburghPA15213 United States

Graph alignment refers to the task of finding the vertex correspondence between two correlated graphs of n vertices. Extensive study has been done on polynomial time algorithms for the graph alignment problem under the Erdős–Rényi graph pair model, where the two graphs are Erdős–Rényi graphs with edge probability qu , correlated under certain vertex correspondence. To achieve exact recovery of the correspondence, all existing algorithms at least require the edge correlation coefficient ρu between the two graphs to be non-vanishing as n → ∞. Moreover, it is conjectured that no polynomial-time algorithm can achieve exact recovery under vanishing edge correlation ρu u ≥ n−Θ(1). We identify a local tree structure, which incorporates one layer of user information and one layer of attribute information, and apply the subgraph counting technique to such structures. A polynomial-time algorithm is proposed that recovers the vertex correspondence for most of the vertices, and then refines the output to achieve exact recovery. The consideration of attribute information is motivated by real-world applications like LinkedIn and Twitter, where user attributes like birthplace and education background can aid alignment. © 1963-2012 IEEE.

关键词： Graph algorithms

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：