检索结果-内蒙古大学图书馆

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

作者： Navarro, Madeline Rozada, Sergio Marques, Antonio G. Segarra, Santiago Dept. of Electrical and Computer Engineering Rice University HoustonTX United States Dept. of Signal Theory and Communications Rey Juan Carlos University Madrid Spain

ISBN: (纸本)9798350368741

This work presents a low-rank tensor model for multidimensional Markov chains. A common approach to simplify the dynamical behavior of a Markov chain is to impose low-rankness on the transition probability matrix. Inspired by the success of these matrix techniques, we present low-rank tensors for representing transition probabilities on multi-dimensional state spaces. Through tensor decomposition, we provide a connection between our method and classical probabilistic models. Moreover, our proposed model yields a parsimonious representation with fewer parameters than matrix-based approaches. Unlike these methods, which impose low-rankness uniformly across all states, our tensor method accounts for the multi-dimensionality of the state space. We also propose an optimization-based approach to estimate a Markov model as a low-rank tensor. Our optimization problem can be solved by the alternating direction method of multipliers (ADMM), which enjoys convergence to a stationary solution. We empirically demonstrate that our tensor model estimates Markov chains more efficiently than conventional techniques, requiring both fewer samples and parameters. We perform numerical simulations for both a synthetic low-rank Markov chain and a real-world example with New York City taxi data, showcasing the advantages of multi-dimensionality for modeling state spaces. © 2025 IEEE.

关键词： low-rank tensors Markov model tensor decomposition

来源：评论

学校读者我要写书评

暂无评论

Exploring the Impact of Conditional Deep Convolutional Generative Adversarial Networks in Brain Tumor Image Classification: A Novel Approach

Exploring the Impact of Conditional Deep Convolutional Gener...

引用

2025 International Conference on Pervasive Computational Technologies, ICPCT 2025

作者： Upadhyay, Deepak Upadhyay, Abhay Sharma, Kuj Bihari Dhondiyal, Shiv Ashish Venu, Nookala Graphic Era Hill University Dept. of Computer Science and Engineering Dehradun India Graphic Era Deemed to Be University Dept. of Computer Science and Engineering Dehradun India Centre for Internet of Things Madhav Institute of Technology & Science Deemed University Gwalior India

ISBN: (纸本)9798331508685

In this paper, we explore the impact of conditional Deep Convolutional Generative Adversarial Networks (cDCGANs) with brain tumor image classification and introduce a new way to improve diagnosis accuracy. Shortage of Data/Data Imbalance: In Medical Imaging datasets, it helps in surging the data scarcity and class imbalance with mimicking generation images using cDCGAN. The detailed investigation demonstrates that synthetic data effectively improves classifier performance by acting as a form of augmentation, providing additional training examples for classifiers and serving as ground truth. From the computational efficiency analysis, we see that there is also scalability in training time with respect to dataset size. Consequently, this study presents cDCGANs as an important asset with which to improve diagnostic accuracy through brain tumor categorization in clinical contexts. © 2025 IEEE.

关键词： Generative adversarial networks

来源：评论

学校读者我要写书评

暂无评论

An optimal group-wise non-rigid image registration technique based clinical decision support for dementia prediction

引用

Multimedia Tools and Applications 2025年 1-26页

作者： Mary, F. Sheeja Asanambigai, V. Fred, A. Lenin Assistant Professor Senior Grade School of Computing Dept. of CSE Vel Tech Rangarajan Dr. Sagunthala R & D Institute of Science and Technology Chennai India Computer Science and Engineering Annamalai University Chidambaram India Dept. of CSE Mar Ephraem College of Engineering and Technology Marthandam India

The fundamental challenge in clinical practice is predicting whether patients with cognitive symptoms or impairment will improve or stay stable. Population ageing is a worldwide phenomena with various impacts. Among dementia, Alzheimer disease (AD) is common in the elderly which is the most common form of dementia and may contribute to 60–70% of cases. An early and precise AD diagnosis enhances patient quality of life. In this study, we propose an optimal group-wise non-rigid image registration technique for dementia (AD) prediction. We introduce an improved cuckoo search optimization (ICSO) algorithm to search optimal group-wise dense correspondence in large image sets in non-rigid image registration which reduce dimensionality of the search space. Then, we illustrate a teacher-student learning based optimization algorithm which cooperates with PLS regression to extract deformation which separates the relationship information suitable for each other from the raw relationship information. After that, we compute the distribution of morphologies of age by fitting smooth percentile curves to these scores using hybrid capsule neural network with least Median Square (hybrid CNN-LMS) curve fitting model. Finally, the simulation results of proposed hybrid CNN-LMS technique are compared to the existing methods to analyze the measures. We observed that the accuracy of our proposed hybrid CNN-LMS technique is 96.23% and 95.82% for RSS and ANDI datasets respectively. © The Author(s), under exclusive licence to Springer science+Business Media, LLC, part of Springer Nature 2025.

关键词： Neurodegenerative diseases

来源：评论

学校读者我要写书评

暂无评论

Online Network Inference from Graph-Stationary Signals with Hidden Nodes

Online Network Inference from Graph-Stationary Signals with ...

引用

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

作者： Buciulea, Andrei Navarro, Madeline Rey, Samuel Segarra, Santiago Marques, Antonio G. Dept. of Signal Theory and Communications Rey Juan Carlos University Madrid Spain Dept. of Electrical and Computer Engineering Rice University HoustonTX United States

ISBN: (纸本)9798350368741

Graph learning is the fundamental task of estimating unknown graph connectivity from available data. Typical approaches assume that not only is all information available simultaneously but also that all nodes can be observed. However, in many real-world scenarios, data can neither be known completely nor obtained all at once. We present a novel method for online graph estimation that accounts for the presence of hidden nodes. We consider signals that are stationary on the underlying graph, which provides a model for the unknown connections to hidden nodes. We then formulate a convex optimization problem for graph learning from streaming, incomplete graph signals. We solve the proposed problem through an efficient proximal gradient algorithm that can run in real-time as data arrives sequentially. Additionally, we provide theoretical conditions under which our online algorithm is similar to batch-wise solutions. Through experimental results on synthetic and real-world data, we demonstrate the viability of our approach for online graph learning in the presence of missing observations. © 2025 IEEE.

关键词： graph signal processing hidden nodes Online graph learning

来源：评论

学校读者我要写书评

暂无评论

Predicting the Likelihood of a Recurrent Febrile Seizure Using Thermographic Images and Deep Learning with Rules Generation 1

Predicting the Likelihood of a Recurrent Febrile Seizure Usi...

引用

1st International Conference on AIML-Applications for engineering and Technology, ICAET 2025

作者： Patil, Pratima Deshpande, Deepa S. MGM University Jawaharlal Nehru Engineering College Dept. of Computer Science and Engineering Chhatrapati Sambhajinagar431003 India

ISBN: (纸本)9798350355611

Recurrent febrile seizures are a common concern in children, often triggered by rapid changes in body temperature during fever episodes. For rapid action and care, early and accurate seizure recurrence prediction is crucial. Limited techniques were developed for predicting the likelihood of a Recurrent Febrile Seizure using a Thermographic image. Several drawbacks were considered for recurrent febrile seizure prediction such as lower performance, high time consumption, and error rate. Hence, the proposed technique is introduced to provide an efficient Deep Learning (DL)-based technique. Initially, the data is gathered from datasets, and the Intra-Attentional Convolutional ResNet (InAtt-CRNet) with the Yolov5-Face model detects the face and reduces the processing time by tacking the face with a bounding box. The patient's forehead or nose temperature is monitored using the Region of Interest (ROI) finding approach following face detection and tracking in each frame. Then, the temperature is predicted using the Enhanced Convolutional Capsule Network (EC2Net). The body temperature, age, duration of fever, and family history are used to construct the clinical decision rule. The Twin-directional Gated Long Short-term memory (TdG-LSTM) is used to predict the likelihood of a subsequent febrile seizure. The proposed technique achieves 98.16% accuracy, 98.45% precision, 0.12 RMSE, and 29 seconds as processing time. © 2025 IEEE.

关键词： Long short-term memory

来源：评论

学校读者我要写书评

暂无评论

COMPACT: Content-aware Multipath Live Video Streaming for Online Classes using Video Tiles 25

COMPACT: Content-aware Multipath Live Video Streaming for On...

引用

16th ACM Multimedia Systems Conference, MMSys 2025

作者： Chaudhary, Shubham Mishra, Navneet Gambhir, Keshav Rajore, Tanmay Bhattacharya, Arani Maity, Mukulika Dept. of Computer Science & Engineering IIIT Delhi Delhi New Delhi India

ISBN: (纸本)9798400714672

The growing popularity of live online classes, even in remote areas, stresses the need for a good and seamless quality of experience to enhance learning. However, these bandwidth-hungry applications challenge the current cellular networks to maintain consistent bandwidth and latency. In this work, we, therefore, propose using the collaboration of multiple devices with their individual cellular networks to support such live video streaming. We design a content-aware system Compact that splits video into foreground and background using video tiles (independently encoded spatial blocks) and streams them over different paths. Compact depends on its scheduler, which exhaustively searches for the best quality based on the network estimates. We extensively evaluate our system using network traces while walking and traveling on the bus or car. Compared to the single path, Compact manages to reduce the median stall and E2E lag by 70.6% and 28.57%, and the tail stall and lag by 83.9% and ≈ 80% on a bus trace. Furthermore, we performed a live experiment to test Compact on the actual cellular network. © 2025 Copyright is held by the owner/author(s). Publication rights licensed to ACM.

关键词： Mobile telecommunication systems

来源：评论

学校读者我要写书评

暂无评论

Activity Prediction for Localizing the Events in Imagined Speech EEG Signals 13

Activity Prediction for Localizing the Events in Imagined Sp...

引用

13th International Winter Conference on Brain-computer Interface, BCI 2025

作者： Balasubramanian, Arun Pandey, Kartik Veer, Gautam Samanta, Debasis Dept. of Computer Science and Enginnering Indian Institute of Technology Kharagpur India

ISBN: (纸本)9798331521929

Imagined speech electroencephalogram (EEG) signals are often collected for longer durations than necessary, leading to a difficulty in understanding the generation of EEG during the task as it is likely that most of the data collected is that of resting state. Developing a filter to identify the segments of the trials with actual information and remove those that contain the EEG from resting states could significantly advance our understanding of EEG signals in neuroscience and biomedical engineering. This work uses derivatives-based features to form the feature vectors used to train the classifiers. Based on feature importance analysis, it has been found that the derivative-based features contributed more to the classification than the traditionally preferred feature, band power in the alpha frequency band. Using the proposed features, the classification performance experienced a significant enhancement, surpassing the results reported in previous studies. The greater accuracy of the classifiers with the proposed features implies that they are effective at filtering out the resting state segments from imagined speech EEG signals. © 2025 IEEE.

关键词： Electroencephalography

来源：评论

学校读者我要写书评

暂无评论

Energy Recovery and Power Management in EVs Using Supercapacitors and Regenerative Braking 2

Energy Recovery and Power Management in EVs Using Supercapac...

引用

2nd International Conference on Advanced Innovations in Smart Cities, ICAISC 2025

作者： Shib, Sowrov Komar Dipto, Durjoy Roy Rahman, Md Tanvir Islam, Md Ashikul Islam, Md Mahafuzul Shufian, Abu American International University-Bangladesh Dept. of Electrical and Electronics Engineering Dhaka Bangladesh Dhaka Bangladesh Lamar University Dept. of Electrical and Computer Engineering TX United States

ISBN: (纸本)9798331506995

Regenerative braking (RB) is crucial in enhancing the efficiency and sustainability of electric vehicles (EVs) by converting kinetic energy into usable electric energy during braking. However, current systems face challenges in maximizing energy recovery and managing power effectively. This research presents an optimized regenerative braking system that combines the benefits of supercapacitors and Brushless DC (BLDC) motors to improve energy recovery efficiency. The study introduces advanced control algorithms for BLDC motors to ensure precise and reliable operation across various driving conditions. Furthermore, refined power management strategies are developed to optimize energy storage between supercapacitors and batteries, enhancing overall system efficiency. The proposed model employs modern control techniques and power management solutions, allowing for real-time energy optimization during braking. With an overall energy recovery efficiency of 92.5%, the system shows significant promise in minimizing energy losses and supporting the sustainability goals of electric vehicles. © 2025 IEEE.

关键词： Battery management systems

来源：评论

学校读者我要写书评

暂无评论

An IoT-Based Monitoring System for Detecting Slope Movement

An IoT-Based Monitoring System for Detecting Slope Movement

引用

Geotechnical Frontiers 2025: Geotechnics of Natural Hazards

作者： Alam, Md. Jobair Bin Manzano, Luis Salgado Debnath, Rahul Ahmed, Ahmed Abdelmoamen Williams, Synia Hartsfield, Markus Dept. of Civil and Environmental Engineering Prairie View A&M Univ. TX United States Dept. of Computer Science TX United States Dept. of Civil and Environmental Engineering TX United States

ISBN: (数字)9780784485996

Shallow landslides pose significant hazards to communities, infrastructure, and natural ecosystems, particularly in regions with steep slopes and vulnerable soil conditions. The slope movement naturally occurs slower depending on environmental conditions and man-made activities and is often triggered by climatic conditions such as precipitation and temperature fluctuations. The initial movement of the slope remains unsighted as the overall earthen slopes’ performance seems satisfactory. However, the initial movement of the slope eventually causes major failure. Therefore, it is crucial to monitor and detect slope movements in real-time, particularly for slopes prone to failure or steep, and timely actions should be taken for the durability and resilience of the earthen slopes. However, monitoring each slope in real-time is a challenging task. This study aimed to develop a distributed sensing system, including slope movement and hydrologic responses executing on the Internet of Things (IoT) at a prototype slope. The proposed system uses various IoT sensors, including tilt sensors and hydrologic sensors, such as moisture sensors and tensiometers, to build a laboratory prototype embankment (0.55 m length, 0.31 m width, and 0.31 m height). The embankment prototype was constructed with Fat Clay (CH). An IoT-integrated data processing and computational system was developed. The experimental results showed that our system could detect negligible movement rotation (0.03°) of slopes. The study’s outcome indicated that our IoT-based slope monitoring system can be cheap and implementable for failure-prone slopes in the field for an effective monitoring system. © ASCE.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Enhancing Brain Tumor Identification through Convolutional Neural Networks and Transformer Models 15

Enhancing Brain Tumor Identification through Convolutional N...

引用

15th IEEE Annual Computing and Communication Workshop and Conference, CCWC 2025

作者： Dudam, Nerajdattu Chato, Lina Motupalli, Nagamani Kashyap, Kristipati Thoyajaksha Sane, Madhusree Bollina, Mounika University of South Dakota Dept. Computer Science Vermillion United States

ISBN: (纸本)9798331507695

An automatic diagnosis of brain tumors based on Deep Learning (DL) is a promising research field to enhance the diagnosis and prognosis of brain tumors. This paper presents an exploration into the use of Convolutional Neural Networks (CNNs) and Vision Transformers (ViT) for brain tumor classification, specifically identifying glioma, meningioma, pituitary tumors, and no tumor from brain Magnetic Reasons Imaging (MRI) scans. We implemented different approaches, such as 1) developing a custom CNN model from scratch, 2) leveraging pre-trained models based on Transfer Learning (TL) (including VGG16 and ResNet50), 3) developing a custom ViT model from scratch, and 4) leveraging pre-trained model in vision transformers, such as ViT B16. A publicly available labeled dataset is used to develop and test our proposed classification Deep Learning models. Our findings reveal that both custom-trained and pre-trained CNN and ViT models achieve high accuracy. The model ViT B16 based on transferred learning achieves the best accuracy of 97.6%, which is slightly better than the CNN models. To enhance the accuracy of the prediction model and reduce the misclassifications, ensemble model based on the combination of ViT B16 and CNN model;this approach increases the accuracy to 98%. © 2025 IEEE.

关键词： Convolutional neural networks

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：