检索结果-内蒙古大学图书馆

2025 IEEE/CVF Winter Conference on Applications of computer Vision, WACV 2025

作者： Liao, Guiqiu Jogan, Matjaz Koushik, Sai Eaton, Eric Hashimoto, Daniel A. University of Pennsylvania Pcaso Laboratory Department of Surgery United States University of Pennsylvania Department of Electrical and Systems Engineering United States University of Pennsylvania Department of Computer and Information Science United States

ISBN: (纸本)9798331510831

Weakly supervised video object segmentation (WSVOS) enables the identification of segmentation maps without requiring extensive annotations of object masks, relying instead on coarse video labels indicating object presence. WSVOS in surgical videos is, however, more challenging due to the complex interaction of multiple transient objects, such as surgical tools moving in and out of the surgical field. In this scenario, state-of-the-art WSVOS methods struggle to learn accurate segmentation maps. We address this problem by introducing ViDeo Spatio-Temporal disentanglement Networks (VDST-Net), a framework to disentangle complex spatio-temporal object interactions using semi-decoupled knowledge distillation to predict high-quality class activation maps (CAMs). A teacher network is designed to help a temporal-reasoning student network resolve activation conflicts, as the student leverages temporal dependencies when specifics about object location and timing in the video are not provided. We demonstrate the efficacy of our framework on a challenging surgical video dataset where objects are, on average, present in less than 60% of annotated frames, and compare our method to state-of-the-art methods on surgical data and on a public dataset commonly used to benchmark WSVOS. Our method outperforms state-of-the-art techniques and generates accurate segmentation masks under video-level weak supervision. Our code is available at: https://***/PCASOlab/VDST-net. © 2025 IEEE.

关键词： Image segmentation

来源：评论

学校读者我要写书评

暂无评论

Privacy-Preserving X-ray Image Enhancement: A GAN-Cybersecurity-Based Approach

Privacy-Preserving X-ray Image Enhancement: A GAN-Cybersecur...

引用

2024 IEEE International Conference on Consumer Electronics, ICCE 2024

作者： Phan, Quoc Bao Nguyen, Linh Nguyen, Tuy Tan Nguyen, Dinh C. Northern Arizona University School of Informatics Computing and Cyber Systems FlagstaffAZ86011 United States The University of Alabama in Huntsville Department of Electrical and Computer Engineering HuntsvilleAL35899 United States

ISBN: (纸本)9798350324136

This paper presents a novel two-stage approach to enhance the quality and privacy of X-ray medical images. The first stage leverages generative adversarial networks (GANs) for effective denoising, eliminating noise and artifacts from X-ray images while improving the visibility of critical anatomical structures. Subsequently, number-Theoretic transform (NTT) polynomial multiplication is integrated with Kyber to accelerate the encryption and decryption of the denoised X-ray images. This encryption safeguards the privacy of sensitive patient data and provides resilience against potential quantum computing attacks, ensuring long-Term data security. Implementing Kyber-based encryption and decryption on a graphics processing unit (GPU) architecture significantly reduces latency, enabling real-Time and secure access to critical healthcare information. © 2024 IEEE.

关键词： GANs Images denoising Kyber number theoretic transform

来源：评论

学校读者我要写书评

暂无评论

Decomposition-Based Classifier Chains for Multi-Dimensional Classification

IEEE Transactions on Artificial Intelligence

引用

IEEE Transactions on Artificial Intelligence 2022年第2期3卷 176-191页

作者： Jia, Bin-Bin Zhang, Min-Ling School of Computer Science and Engineering Southeast University Nanjing210096 China College of Electrical and Information Engineering Lanzhou University of Technology Lanzhou730050 China Key Laboratory of Computer Network and Information Integration Ministry of Education Southeast University Nanjing210096 China

In multi-dimensional classification, the semantics of objects are characterized by multiple class variables from different dimensions. To model the dependencies among class variables, one natural strategy is to build a number of multiclass classifiers in a chaining structure, one per dimension, where the subsequent classifiers on the chain augment the feature space with all labeling information used by the preceding classifiers. However, it is shown that this strategy cannot compete with existing state-of-the-art approaches via comparative studies. One possible reason is that inaccurate predictions of preceding classifiers would degenerate the performance of subsequent ones. Besides, it is more difficult to learn a multiclass classifier than a binary one with the same accuracy, and better performance can be expected if the multi-dimensional classification problem can be solved by building multiple binary classifiers in a chaining structure. Based on these conjectures, this article proposes an approach, which builds a chain of binary classifiers to solve the multi-dimensional classification problem with the help of one-versus-one decomposition. To address the issue that different one-versus-one decomposed problems involve different training examples, the feature space is augmented with the binary predictions of preceding classifiers on the chain to train the subsequent ones. To alleviate the effect of the specified chaining order, the ensemble version of the proposed approach is further investigated. Comparative studies over 20 benchmark datasets clearly show the superiority of the proposed approach against the state-of-the-art multi-dimensional classification baselines. © 2020 IEEE.

关键词： Machine learning

来源：评论

学校读者我要写书评

暂无评论

RUL Estimation for Power Electronic Devices Using RNNs

RUL Estimation for Power Electronic Devices Using RNNs

引用

Prognostics and System Health Management Conference (PHM)

作者： Congrui Cai Zhonghai Lu Embedded Systems Master Program School of Electrical Engineering and Computer Science KTH Royal Institute of Technology Department of Electrical Engineering School of Electrical Engineering and Computer Science KTH Royal Institute of Technology

ISBN: (数字)9798350360585

ISBN: (纸本)9798350360592

Recurrent Neural Networks (RNNs) are commonly used in data-driven approaches to estimate the Remaining Useful Lifetime (RUL) of power electronic devices. RNNs are preferred because their intrinsic feedback mechanisms are better suited to model time-series data. However, the impact of RNN complexity on estimation accuracy is rarely discussed in the literature. This issue is important because choosing a lower-complexity model that delivers the same or similar performance as a higher-complexity model can increase implementation efficiency. In the paper, we use three RNN models, namely, the vanilla version, LSTM (Long Short Term Memory) and GRU (Gated Recurrent Unit) to conduct RUL estimation for power electronic devices. We use two accelerated aging datasets, one dataset targeting the package failure of MOSFETs, and the other dataset targeting package failure of power diodes. Our study shows that a lower-complexity RNN does not necessarily deliver a lower performance. Similarly, a higher-complexity model does not assure a higher performance. As such, our work highlights the importance of selecting a proper neural network for RUL estimation not biased towards complex models. This is especially useful and important for implementing such RUL estimation techniques in embedded resource-constrained and speed-limited computins platforms.

关键词： Performance evaluation Semiconductor device modeling Recurrent neural networks Computational modeling Estimation Power electronics Data models

来源：评论

学校读者我要写书评

暂无评论

Target Detection in Passive MIMO Radar Networks Based on Moments Space 6

Target Detection in Passive MIMO Radar Networks Based on Mom...

引用

6th International Conference on Information Communication and Signal Processing, ICICSP 2023

作者： Li, Bin Li, Jun Guo, Qinghua Xidian University National Key Laboratory of Radar Signal Processing Xian China School of Electrical Computer and Telecommunications Engineering University of Wollongong Wollongong Australia

ISBN: (纸本)9798350339994

This work deals with the issue of target detection in passive multiple-input and multiple-output (MIMO) radar networks. Compared to distributed detection, centralized detection offers better performance but at the cost of high communication and computational costs. To address this issue, we propose a novel moments space based centralized detection method, where the moments of received signal amplitudes are computed as their features, and a low complexity moments space log-likelihood ratio test is proposed. As only the moments need to be sent to the fusion center, the proposed method enjoys low communication and computational costs. Numerical results demonstrate that the proposed method significantly outperforms existing methods in various scenarios. © 2023 IEEE.

关键词： feature fusion likelihood ratio test moments space passive MIMO radar

来源：评论

学校读者我要写书评

暂无评论

Enhancing Breast Cancer Diagnosis with Multi-Resolution Vision Transformers and Robust Decision-Making

引用

IEEE Access 2025年 13卷 89704-89722页

作者： Sabry, Margo Balaha, Hossam Magdy Ali, Khadiga M. Soliman, Tayseer Hassan A. Gondim, Dibson Ghazal, Mohammed Alghamdi, Norah Saleh El-Baz, Ayman Assiut University Information Systems Department Asyut71511 Egypt University of Louisville J. B. Speed School of Engineering Bioengineering Department LouisvilleKY40292 United States Mansoura University Faculty of Medicine Pathology Department Mansoura35511 Egypt University of Louisville Laboratory Medicine Department of Pathology LouisvilleKY40292 United States Abu Dhabi University Electrical Computer and Biomedical Engineering Department Abu Dhabi United Arab Emirates Princess Nourah Bint Abdulrahman University College of Computer and Information Sciences Department of Computer Sciences Riyadh11564 Saudi Arabia

This study aims to improve breast cancer (BC) diagnosis through a novel multi-resolution Vision Transformer (ViT)-based framework with ensemble decision-making, addressing limitations in traditional single-magnification models. The proposed framework uses multiscale feature extraction at three magnification levels (L0, L1, L2 or 16x, 4x, 2x) to capture both fine-grained and high-level tumor features. A stacking ensemble method combines predictions from ViT models trained at these levels, improving classification robustness. Postprocessing techniques, including region-growing and fast-marching level set algorithms, refine whole-slide image (WSI) prediction and postprocessing quality. Performance was evaluated via metrics such as precision, recall, the F1 score, accuracy, and specificity across 50 trials with perturbed conditions. The framework achieved a top accuracy of 97.08%, with precision and recall above 94%. The suggested stack configuration outperformed individual models and other stacking configurations, demonstrating balanced performance with minimal variability. Statistical analysis highlighted the reliability and consistency of the framework under perturbed conditions. The multi-resolution ViT-based framework significantly improves BC classification by integrating multiscale analysis and ensemble decision-making. Its high accuracy and robustness make it a valuable tool for reducing interobserver variability in digital pathology workflows. © 2013 IEEE.

关键词： Image segmentation

来源：评论

学校读者我要写书评

暂无评论

A Proof of Concept Resource Management Scheme for Augmented Reality Applications in 5G systems

arXiv

引用

arXiv 2025年

作者： Nikolaidis, Panagiotis Mostafavi, Samie Gross, James Baras, John Department of Electrical and Computer Engineering The Institute for Systems Research University of Maryland United States School of Electrical Engineering and Computer Science KTH Royal Institute of Technology Sweden

Augmented reality applications are bitrate intensive, delay-sensitive, and computationally demanding. To support them, mobile edge computing systems need to carefully manage both their networking and computing resources. To this end, we present a proof of concept resource management scheme that adapts the bandwidth at the base station and the GPU frequency at the edge to efficiently fulfill roundtrip delay constrains. Resource adaptation is performed using a Multi-Armed Bandit algorithm that accounts for the monotonic relationship between allocated resources and performance. We evaluate our scheme by experimentation on an OpenAirInterface 5G testbed where the considered application is OpenRTiST. The results indicate that our resource management scheme can substantially reduce both bandwidth usage and power consumption while delivering high quality of service. Overall, this work demonstrates that intelligent resource control can potentially establish systems that are not only more efficient but also more sustainable. © 2025, CC BY.

关键词： Quality of service

来源：评论

学校读者我要写书评

暂无评论

Switching Performance Evaluation of two State-of-the-Art SiC MOSFETs

Switching Performance Evaluation of two State-of-the-Art SiC...

引用

Australasian Universities Power engineering Conference, AUPEC

作者： Fredrick Nkado John Fletcher Hang Zhou Scott Tyo Fatemeh Babaeian School of Electrical Engineering and Telecommunications UNSW Sydney Department of Electrical and Computer Systems Engineering Monash University Melbourne

ISBN: (数字)9798350377941

ISBN: (纸本)9798350377958

This paper compares the switching performance of two state-of-the-art Silicon Carbide (SiC) Metal-Oxide-Semiconductor Field-Effect Transistors (MOSFETs) utilising the Double Pulse Test method. The evaluation focuses on assessing the impact of key parameters such as total gate charge $\left(Q_{g}\right)$ and on-resistance $\left(R_{O N}\right)$ on the switching characteristics of the MOSFETs. Through experimental investigation, the switching behaviours, including turn-on and turn-off times, switching losses, and body-diode reverse recovery characteristics ($Q_{r r}$), are analysed and compared between the two devices. The results demonstrate that a SiC MOSFET with a lower total gate charge and higher on-resistance exhibits superior switching performance compared to a SiC MOSFET with a higher total gate charge and lower on-resistance. The study concludes that prioritising low total gate charge for faster switching for pulsed power applications is more critical than minimising conduction losses through a lower $R_{O N}$.

关键词： Performance evaluation MOSFET Power engineering Silicon carbide Switching loss Switches Logic gates Transistors Switching circuits

来源：评论

学校读者我要写书评

暂无评论

Depression Detection in Social Media Posts Using Transformer-Based Models and Auxiliary Features

arXiv

引用

arXiv 2024年

作者： Kerasiotis, Marios Ilias, Loukas Askounis, Dimitris The Decision Support Systems Laboratory School of Electrical and Computer Engineering National Technical University of Athens Athens15780 Greece

The detection of depression in social media posts is crucial due to the increasing prevalence of mental health issues. Traditional machine learning algorithms often fail to capture intricate textual patterns, limiting their effectiveness in identifying depression. Existing studies have explored various approaches to this problem but often fall short in terms of accuracy and robustness. To address these limitations, this research proposes a neural network architecture leveraging transformer-based models combined with metadata and linguistic markers. The study employs DistilBERT, extracting information from the last four layers of the transformer, applying learned weights, and averaging them to create a rich representation of the input text. This representation, augmented by metadata and linguistic markers, enhances the model's comprehension of each post. Dropout layers prevent overfitting, and a Multilayer Perceptron (MLP) is used for final classification. Data augmentation techniques, inspired by the Easy Data Augmentation (EDA) methods, are also employed to improve model performance. Using BERT, random insertion and substitution of phrases generate additional training data, focusing on balancing the dataset by augmenting underrepresented classes. The proposed model achieves weighted Precision, Recall, and F1-scores of 84.26%, 84.18%, and 84.15%, respectively. The augmentation techniques significantly enhance model performance, increasing the weighted F1-score from 72.59% to 84.15%. © 2024, CC BY-NC-ND.

关键词： Linguistics

来源：评论

学校读者我要写书评

暂无评论

Learning Stereo Depth Estimation with Bio-Inspired Spike Cameras

Learning Stereo Depth Estimation with Bio-Inspired Spike Cam...

引用

2022 IEEE International Conference on Multimedia and Expo, ICME 2022

作者： Wang, Yixuan Li, Jianing Zhu, Lin Xiang, Xijie Huang, Tiejun Tian, Yonghong School of Electronic and Computer Engineering Peking University China School of Computer Science Peking University China Department of Electrical and Computer Engineering Shanghai Jiao Tong University China Peng Cheng National Laboratory China

ISBN: (数字)9781665485630

ISBN: (纸本)9781665485630

Bio-inspired spike cameras, offering high temporal resolution spike streams, have brought a new perspective to address common challenges (e.g.,high-speed motion blur) in depth estimation tasks. In this paper, we propose a novel problem setting, spike-based stereo depth estimation, which is the first trail that explores an end-to-end network to learn stereo depth estimation with transformers for spike cameras, named Spike-based Stereo Depth Estimation Transformer (SSDEFormer). We first build a hybrid camera platform and provide a new stereo depth estimation dataset (i.e.,PKU-Spike-Stereo) with spatiotemporal synchronized labels. Then, we propose a novel spike representation to effectively exploit spatiotemporal information from spike streams. Finally, a transformer-based network is designed to generate dense depth maps without a fixed-disparity cost volume. Empirically, it shows that our approach is extremely effective on both synthetic and real-world datasets. The results verify that spike cameras can perform robust depth estimation even in cases where conventional cameras and event cameras fail in fast motion scenarios. © 2022 IEEE.

关键词： Cameras

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：