检索结果-内蒙古大学图书馆

27th International Conference on Discovery Science

作者： van Weezel, Thijs van Ree, Famke Bos, Tychon Bastiaanssen, Patrick Hess, Sibylle Eindhoven Univ Technol Math & Comp Sci Dept Eindhoven Netherlands

ISBN: (纸本)9783031789793;9783031789809

One of the most prominent security challenges to neural networks are adversarial examples - inputs with often barely perceptible perturbations causing misclassification. In this study, we propose a defense mechanism that uses an autoencoder to restore adversarial examples before classification. That is, the autoencoder purifies input data points from potential adversarial perturbations. The method is titled autoencoder-based Adversarial Purification (AAP). We demonstrate the effectiveness of AAP on multiple datasets, attack methods, and perturbation levels. While certain limitations exist, this research offers valuable insights and a promising direction for robust defense mechanisms in adversarial deep learning.

关键词： adversarial attack autoencoder purification

来源：评论

学校读者我要写书评

暂无评论

A Novel Approach to Detection of COVID-19 and Other Respiratory Diseases Using autoencoder and LSTM

引用

SN Computer Science 2025年第1期6卷 1-11页

作者： Malviya, Anjali Dixit, Rahul Shukla, Anupam Kushwaha, Nagendra Indian Institute of Information Technology Pune India Sardar Vallabhbhai National Institute of Technology Surat India

Innumerable approaches of deep learning-based COVID-19 detection systems have been suggested by researchers in the recent past, due to their ability to process high-dimensional, complex data, leading to more accurate prediction of the COVID-19 infected patients. There is a visible dominance of Convolutional Neural Network (CNN) based models analysing chest images like X-rays and Computed Tomography (CT) scans for prediction, while the utilization of audio data for the same is less prevalent. Considering the respiratory system is one of the primary means by which the SARS-CoV-2 virus spreads, respiratory sounds are a potential biomarker for determining the presence of COVID-19. In this paper, we propose a novel approach for the detection of COVID-19 from amidst a dataset comprising of respiratory sound samples of healthy, COVID-19, and other lung diseases which are often misinterpreted as COVID-19. The approach employs an autoencoder for anomaly detection and a Long Short-Term Memory (LSTM) network for the detection of COVID-19 from amongst other lung diseases. The first stage of the model comprises an encoder-decoder-based autoencoder model with baseline reconstruction error, trained in an unsupervised environment, to reconstruct “healthy” audio signals. An LSTM based multi-class classifier is proposed for the second stage to classify the infected samples into the five classes: COVID-19, Bronchiolitis, COPD, Pneumonia and URTI. The experimental results demonstrate the efficacy of our proposed approach in detecting COVID-19 from a 5-class test set of audio samples of patients suffering from respiratory disease, with an accuracy of 98.7%, and an AUC of 1. © The Author(s), under exclusive licence to Springer Nature Singapore Pte Ltd. 2024.

关键词： autoencoder COVID-19 Deep learning Long short-term memory (LSTM) Respiratory-diseases classification

来源：评论

学校读者我要写书评

暂无评论

Reconstruction of EEG and ECG from Single Channel Mixture using Branched autoencoder based Separable Representations

Reconstruction of EEG and ECG from Single Channel Mixture us...

引用

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

作者： Datta, Shreyasi Gubbi, Jayavardhana Pal, Arpan TCS Research India

ISBN: (纸本)9798350368741

The growing use of wearable devices requires accurate and compact representations of high dimensional physiological signals. This work presents a UNet inspired autoencoder to represent and reconstruct multiple neuro-physiological signals from single channel data. The architecture comprises single-encoder/dual-branched decoders to obtain self-attention enabled compact embeddings of mixed ExG (EEG/ECG) signals through decaying encoder-decoder skip connections, for improved representation capability. The embeddings are separable into individual ExG components enabling simultaneous reconstruction of high fidelity EEG and ECG sources. The pretrained encoder can be used for a complex downstream task with minimum fine-tuning. Using the proposed method on a large corpus of single-channel mixed ExG generated from overnight Polysomnography (PSG) recordings, we show subject- and class- independent EEG/ECG reconstructions validated by multiple domain-specific metrics, and evaluate the classification performance of the encoded EEG embeddings into five sleep stages as a downstream task. © 2025 IEEE.

关键词： attention autoencoder ECG EEG reconstruction representation

来源：评论

学校读者我要写书评

暂无评论

Robust Seizure Prediction Based on Riemannian Manifold Enhanced Denoising Adversarial autoencoder

Robust Seizure Prediction Based on Riemannian Manifold Enhan...

引用

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

作者： Peng, Peizhen Wu, Yansheng Yang, Wanqi Wei, Wenxin Yang, Ming School of Computer and Electronic Information The School of Artificial Intelligence Nanjing Normal University Nanjing China

ISBN: (纸本)9798350368741

The seizure early warning devices based on multichannel EEG signals is one of the most used assisted-living strategies for drug-resistant epileptic patients. One of the challenges in the development of these devices is that existing algorithms cannot avoid the effects of electrode loosening. To alleviate such problem, a seizure prediction model robust to corrupted EEG recordings is proposed in this paper. This method intends to learn a stable feature space adapted to different corruption versions via constructing a jointly-optimized autoencoder. Robust information is captured during the reconstruction process by developing a corruption module, while an adversarial training procedure circumvents overfitting using variational inference. Moreover, a regularization term is designed to minimize the distance among various corruption versions in the embedding space with the Riemannian manifold-based distribution alignment. Both raw EEG data and corrupted EEG data are utilized to evaluate prediction performance. Experimental results indicate that this method can predict seizures effectively and stably under the condition of electrode looseness. © 2025 IEEE.

关键词： autoencoder Electroencephalography(EEG) Riemannian manifold seizure prediction time series classification

来源：评论

学校读者我要写书评

暂无评论

Lightweight Botnet Detection for IoT: autoencoder-Based Approach Using Packet Length 3

Lightweight Botnet Detection for IoT: Autoencoder-Based Appr...

引用

3rd International Conference on Intelligent Systems, Advanced Computing, and Communication, ISACC 2025

作者： Aljufri, Muhammad Arifuddin Ahmad, Tohari Aidiel, Muhammad Institut Teknologi Sepuluh Nopember Department of Informatics Surabaya Indonesia

ISBN: (纸本)9798331523893

The growing prevalence of botnet attacks poses a significant threat, particularly as the demand for IoT devices continues to rise. This underscores the need for effective botnet detection techniques tailored for IoT environments. Machine Learning and Deep Learning are often used to tackle this problem with either one showing there are trade-offs between detection and computational performances. To address these challenges, this paper investigates the use of packet length as a key feature for botnet detection, leveraging an autoencoder model. This approach involves feature engineering of packet length and employs an autoencoder to identify anomalous network behavior, making it an efficient and effective solution for detecting Mirai botnet. This method ensures a lightweight detection framework, making it suitable for deployment on resource-constrained IoT devices while maintaining high detection with Precision at 99.64% and F1 Scores at 98,17% respectively tested on the MedBiot dataset. © 2025 IEEE.

关键词： autoencoder Botnet Deep Learning Network Infrastructure Network Security

来源：评论

学校读者我要写书评

暂无评论

autoencoder Based Feature Compression for Bandwidth-Constrained Wireless Sensor Networks 22

Autoencoder Based Feature Compression for Bandwidth-Constrai...

引用

22nd IEEE Consumer Communications and Networking Conference, CCNC 2025

作者： Gao, Kang Bhuyan, Amit Kumar Dutta, Hrishikesh Roy, Avirup Lee, Mei-Hua Biswas, Subir Michigan State University Department of Electrical and Computer Engineering East Lansing MI United States Michigan State University Department of Kinesiology East Lansing MI United States

ISBN: (纸本)9798331508050

This paper introduces an Asymmetric autoencoder (AAE)-driven data compression framework for efficient management of energy, bandwidth, and transmitter complexity in a Wireless Sensor Network (WSN). WSNs are often limited by their ability to process and transmit high-dimensional data due to various constraints, including available energy, processing cycles, and transmission capacity. Achieving an application-specific downstream task executed in a remote receiver under the influence of such sensor node constraints is the focus of the proposed methodology. It places a data compression encoder and decoder at the transmitter and the receiver, respectively. The architecture of the AAE's encoder and decoder can be asymmetric, and the degree of asymmetry can be adjusted based on the computation and processing abilities of the transmitting node, the available bandwidth, and the performance requirements of a specific downstream task. In the proposed framework, the encoder and decoder are jointly trained, enabling the system to extract downstream task-related information from one or more time series inputs. Verified for Human Activity Recognition (HAR), the framework demonstrates effective feature compression while maintaining efficient task performance. © 2025 IEEE.

关键词： autoencoder Feature Compression Human Activity Monitoring Wireless Sensor Network

来源：评论

学校读者我要写书评

暂无评论

Computationally Efficient Hyperspectral Image Classification Using Deep autoencoder and Multistage 2 Dimensional Convolutional autoencoder Architectures

引用

Procedia Computer Science 2025年 260卷 1017-1025页

作者： Sneha R. Burnase Poonam N. Sonar Maitreya Shelare Rajiv Gandhi Institute of Technology

Hyper-spectral imaging (HSI) sensor technology enables extensive coverage with spatial, spectral, and temporal flexibility. However, the substantial volume of spectral-spatial information it provides poses significant challenges in data management and extraction. To address these challenges, various advancements have been made in hyperspectral imaging over recent decades. Key issues in hyperspectral image Classification include the large data size, insufficient training samples, particularly the scarcity of labeled samples and computational burden effective feature extraction methods and the appropriate selection of filters for classification are critical for achieving reliable results without data loss. In this paper, two methods named Deep autoencoder (DAE) and Multistage 2D- Convolutional autoencoder (2D-CAE) proposed with simplified and robust architecture which helped in reduced space and time complexity and efficient in dimentionality reduction. Analysed comparison for Support Vector Machine (SVM), K- nearest Neighbour (KNN), HybridSN, Wavelet-CNN is done with proposed methods (DAE) and Multistage 2D- Convolutional autoencoder (2D- CAE), evaluating their accuracy and computational efficiency with respect to Overall Accuracy (OA), Average Accuracy(AA), Kappa(k), FLOPs, Throughput and Test time required for prediction by performing extensive experiments on publically available dataset Indian Pines (IP), Salina (SA) and Pavia University (PU). FLOPs reduction in DAE is 99.97% and 2D-CAE is 99.95%, throughput reduction in DAE is 99.95% and 2D-CAE is 99.98% when compared individually with HybridSN and Wavelet-CNN, showing significant reduction in space complexity.

关键词： autoencoder Classification Convolution Complexity Hyper-spectral

来源：评论

学校读者我要写书评

暂无评论

Convolution autoencoder based Classification Model for Content Based Image Retrieval

引用

Procedia Computer Science 2025年 258卷 1865-1876页

作者： Rishi Rai Dilip Singh Sisodia

The Content Based Image Retrieval (CBIR) has gained significant importance due to surge of online images and its diverse applications. The CBIR system gets an image as input and returns the most analogous and pertinent image from the database. The CBIR has a wide variety of applications ranging from crime prevention to medical diagnosis and an efficient CBIR system may further improve these areas. In CBIR systems images are converted into binary codes for search and retrieval. There are many approaches for converting images into binary codes but most of them could not properly represent the non-linear relationship exist within the images. To overcome this issue many deep learning architectures and models have been proposed for learning the non-linear relationship within an image. The models used may be unsupervised learning models such as autoencoders or supervised learning models like Convolutional Neural Networks (CNN) with their respective advantages and limitations. Therefore, this paper proposes a convolutional autoencoder model for exploiting the benefits of autoencoder and CNN models for CBIR system. In this way both the data and the class distribution are considered while producing the binary code for CBIR system. The proposed model is evaluated on CIFAR-10 training dataset considering ten classes and performance is also compared with the other existing models. The convolutional autoencoder model has produced better bit codes with more than 95% accuracy as compared to an autoencoder model and has outperformed state of the state of art autoencoder models used for CBIR.

关键词： autoencoder Binary codes CBIR CNN Deep Learning Image Classification

来源：评论

学校读者我要写书评

暂无评论

Convolutional autoencoder-based deep learning for intracerebral hemorrhage classification using brain CT images

引用

Cognitive Neurodynamics 2025年第1期19卷 1-13页

作者： Nageswara Rao, B. Acharya, U. Rajendra Tan, Ru-San Dash, Pratyusa Mohapatra, Manoranjan Sabut, Sukanta Sensing and Computing Lab School of Electronics Engineering Kalinga Institute of Industrial Technology Bhubaneswar India School of Mathematics Physics and Computing University of Southern Queensland Springfield Central QLD Australia National Heart Centre Singapore Singapore Singapore Duke-NUS Medical School Singapore Singapore Department of Computer Science and Engineering Heritage Institute of Technology Kolkata India Department of Radio-Diagnosis Kalinga Institute of Medical Science Bhubaneswar India

Intracerebral haemorrhage (ICH) is a common form of stroke that affects millions of people worldwide. The incidence is associated with a high rate of mortality and morbidity. Accurate diagnosis using brain non-contrast computed tomography (NCCT) is crucial for decision-making on potentially life-saving surgery. Limited access to expert readers and inter-observer variability imposes barriers to timeous and accurate ICH diagnosis. We proposed a hybrid deep learning model for automated ICH diagnosis using NCCT images, which comprises a convolutional autoencoder (CAE) to extract features with reduced data dimensionality and a dense neural network (DNN) for classification. In order to ensure that the model generalizes to new data, we trained it using tenfold cross-validation and holdout methods. Principal component analysis (PCA) based dimensionality reduction and classification is systematically implemented for comparison. The study dataset comprises 1645 (“ICH” class) and 1648 (“Normal” class belongs to patients with non-hemorrhagic stroke) labelled images obtained from 108 patients, who had undergone CT examination on a 64-slice computed tomography scanner at Kalinga Institute of Medical Sciences between 2020 and 2023. Our developed CAE-DNN hybrid model attained 99.84% accuracy, 99.69% sensitivity, 100% specificity, 100% precision, and 99.84% F1-score, which outperformed the comparator PCA-DNN model as well as the published results in the literature. In addition, using saliency maps, our CAE-DNN model can highlight areas on the images that are closely correlated with regions of ICH, which have been manually contoured by expert readers. The CAE-DNN model demonstrates the proof-of-concept for accurate ICH detection and localization, which can potentially be implemented to prioritize the treatment using NCCT images in clinical settings. © The Author(s), under exclusive licence to Springer Nature B.V. 2025.

关键词： autoencoder CAE-DNN Deep learning ICH detection Intracerebral hemorrhage (ICH) Non-contrast CT

来源：评论

学校读者我要写书评

暂无评论

ECG classification using a CNN stacked with an autoencoder trained with synthetic signals

引用

Neural Computing and Applications 2025年 1-21页

作者： Quezada-Prospero, Enrique Mújica-Vargas, Dante Luna-Álvarez, Antonio Rincón, Viridiana Vela Arenas Muñiz, Andrés A. Ciencias Computacionales Centro Nacional de Investigación y Desarrollo Tecnológico Interior del Internado Morelos Cuernavaca 62493 Mexico División de Estudios de Posgrado e Investigación Instituto Tecnológico de Chilpancingo Av. José Francisco Ruiz Massieu No. 5 Guerrero Chilpancingo 39090 Mexico

The detection of cardiac arrhythmias is crucial for preventing severe cardiovascular diseases. As a result, deep learning algorithms have been developed in the state-of-the-art for the automatic detection of these conditions. While these models have produced promising results, they rely on having sufficient data to generalize well. This presents a challenge in the medical field, where data scarcity is often an issue. To address these challenges, this work proposes a hybrid architecture combining a deep autoencoder with a 1D CNN, trained on synthetic signals and evaluated on real signals. The objective is to demonstrate that synthetic signals can complement or even replace real signals when there are insufficient data. The MIT-BIH arrhythmia database, containing five classes, was used for this process. The dataset was split to train a generative adversarial network for generating synthetic signals, while the remaining data were used to evaluate the model on real signals. Signal preprocessing involved transforming the ECG data into 2D representations using different methods, including similarity maps, Gramian angular gields, Markov transition fields, and co-occurrence matrices (CoMs), each of which was evaluated with the proposed model to observe the performance of each one. The extracted features were compressed by an autoencoder and then passed through a 1D CNN for classification. The best results in real data evaluation were obtained with CoMs, achieving an accuracy of 99.7%, precision of 99.7%, recall of 99.7%, specificity of 99.9%, F1-score of 99.7%, Matthews correlation coefficient of 99.6%, and Cohen’s kappa of 99.3%. © The Author(s), under exclusive licence to Springer-Verlag London Ltd., part of Springer Nature 2025.

关键词： 2D signal transformations autoencoder ECG classification Generative adversarial network

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：