检索结果-内蒙古大学图书馆

GAIT based human Parkinson’s disease detection using fused features with multi-kernel support vector machine

International Journal of Information Technology (Singapore) 2025年第3期17卷 1387-1395页

作者： Jhapate, Arun Kumar Shrivastava, Hemang Department of Computer Science and Engineering Sagar Institute of Research and Technology Madhya Pradesh Bhopal India Institute of Advance Computing SAGE University Madhya Pradesh Indore India

Gait analysis plays a pivotal role in the detection and monitoring of Parkinson’s disease (PD), offering insights into motor impairments characteristic of the condition. This study proposes a novel approach employing advanced feature extraction and machine learning algorithms to enhance the classification of PD depends on gait features. In this study, both local as well as global features are extracted from the gait images to facilitate a classification process. Local feature extraction is achieved using an optimized convolutional neural network (OCNN), while global feature extraction is done by the Swin Transformer architecture. To further improve feature extraction, an Adaptive Glowworm Swarm Optimization (AGSO) algorithm is applied to improve the performance of the CNN. The fused local and global features are then fed into a Multi-Kernel Support Vector Machine (MK-SVM) classifier, which serves as the core of the proposed model. The performance of the model is evaluated using the GAIT-IT datasets, assessing metrics such as accuracy, precision, recall, and F-measure. As a result of the proposed model, performance is superior to that of traditional methods in terms of accuracy and reliability, offering a promising approach for PD detection based on gait analysis. © Bharati Vidyapeeth's Institute of computer Applications and Management 2024.

关键词： Gait analysis Global features Local features Machine learning Multi kernel SVM Optimized CNN Parkinson’s disease Swin transformer

来源：评论

学校读者我要写书评

暂无评论

Dual-level Mixup for Graph Few-shot Learning with Fewer Tasks 25

Dual-level Mixup for Graph Few-shot Learning with Fewer Task...

引用

34th ACM Web Conference, WWW 2025

作者： Liu, Yonghao Li, Mengyu Giunchiglia, Fausto Huang, Lan Li, Ximing Feng, Xiaoyue Guan, Renchu College of Computer Science and Technology Jilin University Changchun China Department of Information Engineering and Computer Science University of Trento Trento Italy Key Laboratory of Symbolic Computation and Knowledge Engineering of the Ministry of Education China

ISBN: (纸本)9798400712746

Graph neural networks have been demonstrated as a powerful paradigm for effectively learning graph-structured data on the web and mining content from it. Current leading graph models require a large number of labeled samples for training, which unavoidably leads to overfitting in few-shot scenarios. Recent research has sought to alleviate this issue by simultaneously leveraging graph learning and meta-learning paradigms. However, these graph meta-learning models assume the availability of numerous meta-training tasks to learn transferable meta-knowledge. Such assumption may not be feasible in the real world due to the difficulty of constructing tasks and the substantial costs involved. Therefore, we propose a SiMple yet effectIve approach for graph few-shot Learning with fEwer tasks, named SMILE. We introduce a dual-level mixup strategy, encompassing both within-task and across-task mixup, to simultaneously enrich the available nodes and tasks in meta-learning. Moreover, we explicitly leverage the prior information provided by the node degrees in the graph to encode expressive node representations. Theoretically, we demonstrate that SMILE can enhance the model generalization ability. Empirically, SMILE consistently outperforms other competitive models by a large margin across all evaluated datasets with in-domain and cross-domain settings. Our anonymous code can be found here. © 2025 Copyright held by the owner/author(s).

关键词： Graph neural networks

来源：评论

学校读者我要写书评

暂无评论

Adaptive Fine-Grained Feature Mining and RoI Feature Interaction Network for Small Object Detection in Aerial Images

Adaptive Fine-Grained Feature Mining and RoI Feature Interac...

引用

International Conference on Acoustics, Speech, and Signal Processing (ICASSP)

作者： Fangmin Xie Zhan Xiong Zhuoyue Wang Kun Yang Guoqiang Xiao College of Computer and Information Science Southwest University China Institute of Advanced Computer Science Leiden University Netherlands Department of Electrical Engineering and Computer Sciences University of California Berkeley Berkeley USA

ISBN: (数字)9798350368741

ISBN: (纸本)9798350368758

Object detection in drone view images remains challenging due to small-scale objects distributed non-uniformly and exhibiting weak semantic features. Additionally, external factors such as lighting conditions, viewing angles, and background clutter further complicate the detection of small objects. To address these issues, we propose a small object detector called Adaptive Fine-grained Feature Mining Network (AFMNet) based on an anchor-free framework. Firstly, an Adaptive Fine-grained Feature Mining (AFFM) module is designed to generate Regions of Interest (RoI) on high-resolution feature maps and mine fine-grained information that benefits the representation of small objects. Secondly, an Inter-RoI Feature Interaction (IRFI) block is developed to efficiently integrate complementary cues— namely, deep layer semantic information and shallow layer spatial details—and enhance RoI features. Experiments on the VisDrone dataset demonstrate that our method achieves state-of-the-art performance. The source code of our AFMNet is available at https://***/xiefangmin/AFMNet.

关键词： Adaptive systems Source coding Semantics Lighting Object detection Detectors Signal processing Feature extraction Speech processing Drones

来源：评论

学校读者我要写书评

暂无评论

Refractory Poisson: The Effect of the Refractory Period on Neural Rate Coding

引用

IEEE Transactions on Communications 2025年

作者： Nasiraee, Mohammad Kordy, Hussain Montazery Kazemitabar, Javad Babol Noshirvani University of Technology Department of Electrical Computer Engineering Babol Iran

The neuron's membrane potential build-up allows for refractoriness. In this context, the refractory effect is often the primary reason spike trains are not accurately described by traditional counting distributions. Inspired by this biological phenomenon, this paper introduces a counting distribution that includes a parameter specifying the length of the refractory period in neural spike generation. Specifically, the paper considers a lower bound on the intervals between any two spikes that follow an exponential or Gamma distribution. The proposed distribution is a generalized form of standard counting distributions, such as the Poisson distribution, and is named the refractory Poisson distribution. This study aims to provide estimation methods for the proposed distribution. Additionally, it introduces a novel estimation approach called "one-shot"maximum likelihood estimation. The research also provides guidance on the maximum amount of information that can be reliably transferred through a single neuron due to refractoriness. The optimality conditions for the capacity-achieving distribution in the capacity per unit cost problem are presented, which, to the best of our knowledge, have not been reported in the literature. The proposed counting distribution can be used for numbering any random phenomena where a refractory period exists between occurrences. Additionally, the distribution can be applied in various scenarios. For example, it can replace the Poisson distribution in the Generalized Linear Model as an alternative counting distribution. © 2025 IEEE.

关键词： Neurons

来源：评论

学校读者我要写书评

暂无评论

AUDIO VISUAL SEGMENTATION THROUGH TEXT EMBEDDINGS

arXiv

引用

arXiv 2025年

作者： Lee, Kyungbok Zhang, You Duan, Zhiyao Department of Computer Science University of Rochester RochesterNY United States Department of Electrical and Computer Engineering University of Rochester RochesterNY United States

The goal of Audio-Visual Segmentation (AVS) is to localize and segment the sounding source objects from the video frames. Researchers working on AVS suffer from limited datasets because hand-crafted annotation is expensive. Recent works attempt to overcome the challenge of limited data by leveraging the segmentation foundation model, SAM, prompting it with audio to enhance its ability to segment sounding source objects. While this approach alleviates the model’s burden on understanding visual modality by utilizing pre-trained knowledge of SAM, it does not address the fundamental challenge of the limited dataset for learning audio-visual relationships. To address these limitations, we propose AV2T-SAM, a novel framework that bridges audio features with the text embedding space of pre-trained text-prompted SAM. Our method leverages multimodal correspondence learned from rich text-image paired datasets to enhance audio-visual alignment. Furthermore, we introduce a novel feature, fCLIP ☉ fCLAP, which emphasizes shared semantics of audio and visual modalities while filtering irrelevant noise. Experiments on the AVSBench dataset demonstrate state-of-the-art performance on both datasets of AVSBench. Our approach outperforms existing methods by effectively utilizing pretrained segmentation models and cross-modal semantic alignment. © 2025, CC BY.

关键词： Semantics

来源：评论

学校读者我要写书评

暂无评论

Enhanced Body Composition Estimation from 3D Body Scans

Enhanced Body Composition Estimation from 3D Body Scans

引用

18th International Joint Conference on Biomedical engineering Systems and Technologies, BIOSTEC 2025

作者： Feng, Boyuan Zheng, Yijiang Cheng, Ruting Feng, Shuya Vaziri, Khashayar Hahn, James Department of Computer Science George Washington University WashingtonDC United States Department of Computer Science and Engineering University of Connecticut StorrsCT United States Department of Surgery The George Washington University Medical Faculty Associates WashingtonDC United States

Accurate body composition assessment is essential for evaluating health and diagnosing conditions like sar copenia and cardiovascular disease. Approaches for accurately measuring body composition, such as Dual Energy X-ray Absorptiometry (DXA) and Magnetic Resonance Imaging (MRI), are precise but costly and limited in accessibility. Some studies have explored predicting body composition by using shapes since 3D scanning techniques allow for precise and efficient digital measurements of body shape. This study introduces an enhanced method using 3D body scanning integrated with a part-to-global Multilayer Perceptron (MLP) network that incorporates predefined high-level features for body composition prediction. For lean mass es timation, our method achieved a root mean square error (RMSE) of 2.85 kg. For fat mass estimation, the RMSE was 2.50 kg, and for bone mineral content (BMC), the RMSE was 193.50 g. These results represent substantial improvements over existing methods, highlighting the effectiveness and reliability of our approach in accurately predicting body composition metrics. © 2025 by SCITEPRESS– science and Technology Publications, Lda.

关键词： Diagnosis

来源：评论

学校读者我要写书评

暂无评论

Brain MRI Segmentation with Language-Driven Detection and Context-Aware Descriptions

Brain MRI Segmentation with Language-Driven Detection and Co...

引用

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

作者： Fu, Qiang Xia, Xinyuan Hong, Yi Department of Computer Science and Engineering Shanghai Jiao Tong University Shanghai20040 China Shanghai Jiao Tong University Shanghai200240 China

ISBN: (纸本)9798350368741

Brain MRI segmentation is critical for diagnosis and treatment planning, but existing methods are often limited by their task-specific designs and lack of generalizability. A significant challenge lies in integrating multiple brain imaging datasets with varying structural labels for comprehensive analysis. To address this, we propose GroundingSeg, a language-driven framework that unifies segmentation and object detection with phrase grounding to enhance segmentation outcomes. GroundingSeg leverages semantic-aware detection queries and auto-generated descriptions to provide fine-grained and context-aware segmentation of brain structures, using object detection to improve segmentation precision. Our results show that GroundingSeg outperforms existing methods in adaptability and precision, offering a robust solution for brain MRI segmentation. Our source code is available at https://***/ABC67876/GroundingSeg/. © 2025 IEEE.

关键词： Brain MRI Segmentation Language-Driven Segmentation Sematic-aware Detection Queries

来源：评论

学校读者我要写书评

暂无评论

Enabling Robust Security in MQTT-Based IoT Networks with Dynamic Resource-Aware Key Sharing

引用

Procedia computer science 2025年 252卷 633-642页

作者： Sharadadevi Kaganurmath Nagaraj Cholli Department of Computer Science Engineering Global Academy of Technology Bengaluru India Department of Information Science Engineering RV College of Engineering Bengaluru

This paper presents a novel approach aimed at developing a secure secret key-sharing system optimized for resource-constrained Internet of Things (IoT) devices. Focusing on the MQTT protocol, the research endeavors to establish secure communication channels between IoT devices and brokers, thereby enhancing the overall security of MQTT-based IoT deployments. This research introduces a novel Dynamic Lightweight Authentication for MQTT (DLA-MQTT) mechanism designed to the unique needs of IoT devices operating under the MQTT protocol. The DLA-MQTT mechanism leverages an innovative lightweight Generalized Feedback Shift Register (GFSR)-based Pseudo-Random Number Generator (PRNG) to generate ephemeral keys, ensuring secure communication while addressing the limitations of computational power and energy resources. Through a detailed comparative analysis with existing cryptographic solutions, the DLA-MQTT mechanism demonstrates superior performance in terms of computational overhead, energy consumption, and execution time, while maintaining robust security against common attacks such as Man-in-the-Middle (MitM) and Denial of Service (DoS). he proposed algorithm’s adaptability and scalability are validated using the Cooja simulator, where simulated IoT networks are subjected to various threat scenarios. The results confirm the DLA-MQTT mechanism’s efficacy, showcasing a significant reduction in resource utilization without compromising the strength of the security provided.

关键词： Internet of Things(IoT) Message Queuing Telemetry Transport (MQTT) Lightweight Cryptography Resource-Constrained Devices Security Pseudo-Random Number Generator Generalised feedback Shift Register (GFSR) Ephemeral Key Generation

来源：评论

学校读者我要写书评

暂无评论

Homomorphic Evaluation Cluster Architecture for Fully Homomorphic Encryption

IEEE Open Journal of Circuits and Systems

引用

IEEE Open Journal of Circuits and Systems 2025年

作者： Lee, Hanyoung Satriawan, Ardianto Lee, Hanho Inha University Department of Electrical and Computer Engineering Incheon22212 Korea Republic of

Fully Homomorphic Encryption (FHE) allows computational processing of encrypted data on cloud servers, providing high security and enabling safe data utilization. As homomorphic multiplication progresses with encrypted data, noise accumulates, requiring a process called bootstrapping to restore the noise level of the new ciphertext ct′. Bootstrapping involves linear transformation processes, such as Coefficient to Slots and Slots to Coefficient, where most operations used are rotation. Rotation shifts elements in slots to new positions based on rotation index k. However, the computational cost and memory bandwidth required for a rotation adds significant overhead and limits the ability to perform FHE operations. Therefore, an efficient implementation of rotation is crucial for high-performance FHE applications. To address this problem, we optimized the datapath of rotation in the CKKS scheme to be hardware-friendly and proposed a homomorphic evaluation cluster hardware accelerator tailored for FHE workloads. Our architecture is aware of the computational and memory constraints of field programmable gate arrays (FPGAs) and performs number theoretic transform (NTT), its inverse (INTT), key multiplication, base conversion, and automorphism in a single cluster. We implemented our design in the AMD Alveo U280 FPGA platform. With a polynomial length of 216 and operating at 250 MHz as a rotation accelerator, the design implementation on the FPGA shows a speed-up of about 700× compared to the CPU implementation in OpenFHE. Compared to the GPU implementation, it shows a 1.77× speed-up, and compared to previous FPGA implementations, it shows a 1.13× better. © 2020 IEEE.

关键词： Linear transformations

来源：评论

学校读者我要写书评

暂无评论

A Heterogeneous Multiscale Method for Efficient Simulation of Power Systems with Inverter-Based Resources

arXiv

引用

arXiv 2025年

作者： Huang, Kaiyang Xiong, Min Liu, Yang Sun, Kai Department of Electrical Engineering and Computer Science University of Tennessee KnoxvilleTN37996 United States

As inverter-based resources (IBRs) penetrate power systems, the dynamics become more complex, exhibiting multiple timescales, including electromagnetic transient (EMT) dynamics of power electronic controllers and electromechanical dynamics of synchronous generators. Consequently, the power system model becomes highly stiff, posing a challenge for efficient simulation using existing methods that focus on dynamics within a single timescale. This paper proposes a Heterogeneous Multiscale Method for highly efficient multi-timescale simulation of a power system represented by its EMT model. The new method alternates between the microscopic EMT model of the system and an automatically reduced macroscopic model, varying the step size accordingly to achieve significant acceleration while maintaining accuracy in both fast and slow dynamics of interests. It also incorporates a semi-analytical solution method to enable a more adaptive variable-step mechanism. The new simulation method is illustrated using a two-area system and is then tested on a detailed EMT model of the IEEE 39-bus system. Copyright © 2025, The Authors. All rights reserved.

关键词： Electromagnetic transients

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：