检索结果-内蒙古大学图书馆

Studying the Manifold Structure of Alzheimers Disease: A Deep Learning Approach Using convolutional autoencoders

IEEE JOURNAL OF BIOMEDICAL AND HEALTH INFORMATICS 2020年第1期24卷 17-26页

作者： Martinez-Murcia, Francisco J. Ortiz, Andres Gorriz, Juan-Manuel Ramirez, Javier Castillo-Barnes, Diego Univ Malaga Dept Communicat Engn Malaga 29016 Spain Andalusian Res Inst Data Sci & Computat Intellige Granada 18071 Spain Univ Granada Dept Signal Theory Networking & Commun E-18071 Granada Spain DaSCI Granada 18071 Spain

Many classical machine learning techniques have been used to explore Alzheimers disease (AD), evolving from image decomposition techniques such as principal component analysis toward higher complexity, non-linear decomposition algorithms. With the arrival of the deep learning paradigm, it has become possible to extract high-level abstract features directly from MRI images that internally describe the distribution of data in low-dimensional manifolds. In this work, we try a new exploratory data analysis of AD based on deep convolutional autoencoders. We aim at finding links between cognitive symptoms and the underlying neurodegeneration process by fusing the information of neuropsychological test outcomes, diagnoses, and other clinical data with the imaging features extracted solely via a data-driven decomposition of MRI. The distribution of the extracted features in different combinations is then analyzed and visualized using regression and classification analysis, and the influence of each coordinate of the autoencoder manifold over the brain is estimated. The imaging-derived markers could then predict clinical variables with correlations above 0.6 in the case of neuropsychological evaluation variables such as the MMSE or the ADAS11 scores, achieving a classification accuracy over 80 for the diagnosis of AD.

关键词： Feature extraction Magnetic resonance imaging Manifolds Alzheimer's disease Computer architecture Deep learning Alzheimer's disease deep learning convolutional autoencoder manifold learning data fusion

来源：评论

学校读者我要写书评

暂无评论

Feature-enhanced X-ray imaging using fused neural network strategy with designable metasurface

引用

NANOPHOTONICS 2023年第19期12卷 3793-3805页

作者： Shi, Hao Sun, Yuanhe Liang, Zhaofeng Cao, Shuqi Zhang, Lei Zhu, Daming Wu, Yanqing Yao, Zeying Chen, Wenqing Li, Zhenjiang Yang, Shumin Zhao, Jun Wang, Chunpeng Tai, Renzhong Chinese Acad Sci Shanghai Adv Res Inst Shanghai Synchrotron Radiat Facil Shanghai 201204 Peoples R China Chinese Acad Sci Shanghai Inst Appl Phys Shanghai 201800 Peoples R China Univ Chinese Acad Sci Beijing 100049 Peoples R China Nanjing Univ Nanjing 210093 Peoples R China Zhejiang Lab Hangzhou 311101 Peoples R China Chinese Acad Sci Innovat Acad Microsatellites Shanghai 201203 Peoples R China

Scintillation-based X-ray imaging can provide convenient visual observation of absorption contrast by standard digital cameras, which is critical in a variety of science and engineering disciplines. More efficient scintillators and electronic postprocessing derived from neural networks are usually used to improve the quality of obtained images from the perspective of optical imaging and machine vision, respectively. Here, we propose to overcome the intrinsic separation of optical transmission process and electronic calculation process, integrating the imaging and postprocessing into one fused optical-electronic convolutional autoencoder network by affixing a designable optical convolutional metasurface to the scintillator. In this way, the convolutional autoencoder was directly connected to down-conversion process, and the optical information loss and training cost can be decreased simultaneously. We demonstrate that feature-specific enhancement of incoherent images is realized, which can apply to multi-class samples without additional data precollection. Hard X-ray experimental validations reveal the enhancement of textural features and regional features achieved by adjusting the optical metasurface, indicating a signal-to-noise ratio improvement of up to 11.2 dB. We anticipate that our framework will advance the fundamental understanding of X-ray imaging and prove to be useful for number recognition and bioimaging applications.

关键词： X-ray optics metasurface convolutional autoencoder denoising

来源：评论

学校读者我要写书评

暂无评论

An Energy-Efficient Reconfigurable autoencoder Implementation on FPGA 1

引用

Intelligent Systems Conference (IntelliSys)

作者： Isik, Murat Oldland, Matthew Zhou, Lifeng Drexel Univ Elect & Comp Engn Philadelphia PA 19104 USA

ISBN: (数字)9783031477218

ISBN: (纸本)9783031477201;9783031477218

autoencoders are unsupervised neural networks that are used to process and compress input data and then reconstruct the data back to the original data size. This allows autoencoders to be used for different processing applications such as data compression, image classification, image noise reduction, and image coloring. Hardware-wise, re-configurable architectures like Field Programmable Gate Arrays (FPGAs) have been used for accelerating computations from several domains because of their unique combination of flexibility, performance, and power efficiency. In this paper, we look at the different autoencoders available and use the convolutional autoencoder in both FPGA and GPU-based implementations to process noisy static MNIST images. We compare the different results achieved with the FPGA and GPU-based implementations and then discuss the pros and cons of each implementation. The evaluation of the proposed design achieved 80% accuracy and our experimental results show that the proposed accelerator achieves a throughput of 21.12 Giga-Operations Per Second (GOP/s) with a 5.93 W on-chip power consumption at 100 MHz. The comparison results with off-the-shelf devices and recent state-of-the-art implementations illustrate that the proposed accelerator has obvious advantages in terms of energy efficiency and design flexibility. We also discuss future work that can be done with the use of our proposed accelerator.

关键词： convolutional autoencoder Reconfigurable architecture FPGA Energy efficiency

来源：评论

学校读者我要写书评

暂无评论

Acoustic Anomaly Detection of Machinery using autoencoder based Deep Learning 32

Acoustic Anomaly Detection of Machinery using Autoencoder ba...

引用

32nd Southern African Universities Power Engineering Conference (SAUPEC)

作者： Chinnasamy, Mark Damien Sumbwanyambe, Mbuyu Hlalele, Tlotlollo Sidwell Univ South Africa Dept Elect Engn Johannesburg South Africa

ISBN: (纸本)9798350371345

This research paper focuses on the domain of Acoustic Anomaly Detection (AAD) in industrial machinery using Deep Learning techniques. The primary objective of this study is to develop a reliable system for detecting anomalies in the acoustic patterns of industrial machines. In large-scale industrial environments, the early detection of faults and anomalies is critical to ensure the uninterrupted operation of machinery, minimize downtime, and optimize maintenance efforts. The datasets used in this research encompass a range of Signal-to-Noise Ratios (SNR) to simulate diverse operating conditions for industrial fans and pumps. After preprocessing each audio sample was transformed into 9 segments of shape 128 x 32 log mel-spectrograms. The study encompasses a comprehensive analysis of convolutional autoencoders (CAE) and Dense autoencoders (DAE). These models are trained and evaluated against real-world industrial datasets (MIMII dataset), and their performance is meticulously compared. The results for the DAE and CAE had shown over 0.80 at the 6 dB SNR level and decaying results as the SNR level worsens. This research confirms some of the trends pointed out by the literature and provides detailed insight into how the autoencoders are developed and their properties could be used in order to detect anomalous behavior in audio data.

关键词： autoencoder Acoustic Detection convolutional autoencoder Dense autoencoder ML Unsupervised Anomaly Detection

来源：评论

学校读者我要写书评

暂无评论

AI-Based Temperature Monitoring System for Hydro Generators 23

AI-Based Temperature Monitoring System for Hydro Generators

引用

23rd International Symposium INFOTEH-JAHORINA (INFOTEH)

作者： Milic, Sasa D. Kozicic, Misa Univ Belgrade Elect Engn Inst Nikola Tesla Belgrade Serbia HPPs Derdap EPS JSC Belgrade Kladovo Serbia

ISBN: (纸本)9798350329940;9798350307368

Hydrogenerators operate in challenging environments, and temperature variations can significantly impact their performance. Temperature monitoring systems often rely on remote infrared and contact real-time temperature measurement of the generator. AI-based temperature monitoring systems offer a state-of-the-art solution, providing continuous, automated, and predictive maintenance capabilities. The paper presents a complex AI-based fuzzy decision-making algorithm for anomaly (overheated rotor pole) detection and failure prediction. Time-based analysis of operational parameters from SCADA systems is essential for failure prediction. The use of recurrent neural networks for time-series analysis and convolutional autoencoders for anomaly detection is especially emphasized. The paper also presents a fuzzy decision-making method that uses fuzzy logic to make decisions after analyzing the output of the AI models.

关键词： temperature monitoring system hydro generator fuzzy decision-making LSTM convolutional autoencoder

来源：评论

学校读者我要写书评

暂无评论

R-CAE-Informer Based Short-Term Load Forecasting by Enhancing Feature in Smart Grids 20th

R-CAE-Informer Based Short-Term Load Forecasting by Enhancin...

引用

20th International Conference on Intelligent Computing (ICIC)

作者： Zhang, Yiying Liu, Ke Dong, Yanping Li, Siwei Li, Wenjing Tianjin Univ Sci & Technol Tianjin 300457 Peoples R China Beijing Fibrlink Commun Co LTD Beijing 100070 Peoples R China State GridInformat & Telecommun Co Beijing 100192 Peoples R China

ISBN: (纸本)9789819756650;9789819756667

As renewable energy usage increases, power systems become more intricate and demand fluctuations intensify. Accurate short-term load forecasting (STLF) is vital for balancing energy supply and demand. Traditional models often struggle with long input sequences, risking critical feature loss due to inadequate capture of long-period characteristics. To address these challenges, we introduce the R-VAE-Informer, a novel forecasting approach that combines Deep Residual Networks (DRN) and convolutional autoencoders (CAE). This model leverages the ProbSparse self-attention mechanism and attention distillation techniques to manage the quadratic complexity of long load sequences more efficiently. It also incorporates a Residual convolutional autoencoder (R-CAE) to produce two-dimensional representations of load data, thereby enhancing feature representation and mitigating potential feature loss. Furthermore, the TempoLocode submodule integrates time and positional data, effectively capturing long-term dependencies and periodic variations in time series. Tests on two public datasets demonstrate that our model achieves a Mean Absolute Percentage Error (MAPE) as low as 0.97%, confirming its strong generalization capabilities. Overall, this model markedly improves the precision and sophistication of STLF.

关键词： short-term load forecasting convolutional autoencoder Residual Network deep learning R-VAE-Informer

来源：评论

学校读者我要写书评

暂无评论

Detecting 5G Signal Jammers Using Spectrograms with Supervised and Unsupervised Learning 59

Detecting 5G Signal Jammers Using Spectrograms with Supervis...

引用

59th Annual IEEE International Conference on Communications (IEEE ICC)

作者： Varotto, Matteo Valentin, Stefan Tomasin, Stefano Darmstadt Univ Appl Sci Dept Comp Sci Darmstadt Germany Univ Padua Dept Informat Engn Dept Math Padua Italy

ISBN: (纸本)9798350304060;9798350304053

Cellular networks are potential targets of jamming attacks to disrupt wireless communications. Since the fifth generation (5G) of cellular networks enables mission-critical applications, such as autonomous driving or smart manufacturing, the resulting malfunctions can cause serious damage. This paper proposes to detect broadband jammers by an online classification of spectrograms. These spectrograms are computed from a stream of in-phase and quadrature (IQ) samples of 5G radio signals. We obtain these signals experimentally and describe how to design a suitable dataset for training. Based on this data, we compare two classification methods: a supervised learning model built on a basic convolutional neural network (CNN) and an unsupervised learning model based on a convolutional autoencoder (CAE). After comparing the structure of these models, their performance is assessed in terms of accuracy and computational complexity.

关键词： 5G Wireless Intrusion Detection Jammer convolutional autoencoder convolutional Neural Network Spectrogram Software Defined Radio

来源：评论

学校读者我要写书评

暂无评论

Deep learning in the SKA era: patterns in the SNR population with unsupervised ML methods 8

Deep learning in the SKA era: patterns in the SNR population...

引用

Conference on Software and Cyberinfrastructure for Astronomy VIII

作者： Bufano, F. Bordiu, C. Cecconello, T. Munari, M. Hopkins, A. M. Ingallinera, A. Leto, P. Loru, S. Riggi, S. Sciacca, E. Vizzari, G. Demarco, A. Buemi, C. S. Cavallaro, F. Trigilio, C. Umana, G. INAF Osservatorio Astrofis Catania Via Santa Sofia 78 I-95123 Catania Italy Univ Catania Dept Elect Elect & Comp Engn Viale Andrea Doria 6 I-95125 Catania Italy Univ Milano Bicocca Viale Sarca 336 I-20126 Milan Italy Macquarie Univ Sch Math & Phys Sci 12 Wallys Walk Sydney NSW 2109 Australia Univ Malta Inst Space Sci & Astron Maths & Phys Bldg MSD-2080 Msida Malta

ISBN: (纸本)9781510675261;9781510675254

The Square Kilometre Array precursors are releasing the first data of their large-field continuum surveys. The complexity of such datasets makes clear that deep learning is the primary solution for handling an overwhelming volume of data also in the radio astronomy field. Within this framework, our research group is taking a forefront position in various research initiatives aimed at assessing the effectiveness of ML techniques on survey data from ASKAP and MeerKAT. In this work we show how an unsupervised multi-stage pipeline is able to discover physically meaningful clusters within the heterogeneous Supernova Remnant (SNR) population: a convolutional autoencoder extracts features from multiwavelength imagery of a SNR sample;then an unsupervised clustering process operates on the latent space to identify patterns. Despite a large number of outliers, we were able to find a new classification system, in which most clusters relate to the presence of certain features regarding not only the morphology but also the relative weight of the different frequencies.

关键词： Machine Learning Astrophysics convolutional autoencoder Supernova Remnant

来源：评论

学校读者我要写书评

暂无评论

A novel unsupervised deep learning method for the generalization of urban form

引用

Geo-Spatial Information Science 2022年第4期25卷 568-587页

作者： Jihong Cai Yimin Chen School of Geography and Planning Sun Yat-sen UniversityGuangzhouChina

Accurate delineation of urban form is essential to understand the impacts that urbanization has on the environment and regional *** supervised classification of urban form requires a rigidly defined scheme and high-quality sample data with class *** to the complexity of urban systems,it is challenging to consistently define urban form types and collect metadata to describe ***,in this study,we propose a novel unsupervised deep learning method for urban form delineation while avoiding the limitations of conventional super-vised urban form classification *** novelty of the proposed method is the Multiscale Residual convolutional autoencoder(MRCAE),which can learn the latent representation of differ-ent urban form *** vectors can be further used to generalize urban form types by using Self-Organizing Map(SOM)and the Gaussian Mixture Model(GMM).The proposed method is applied in the metropolitan area of Guangzhou-Foshan,*** MRCAE model along with SOM and GMM is used to generalize the urban form types from satellite *** physical and functional properties of each urban form type are also analyzed using several auxiliary datasets,including building footprints,Points-of-Interests(POIs)and Tencent User Density(TUD)*** results reveal that the urban form map generated based on the MRCAE can explain 55%of the building height distribution and 55%of the building area distribution,which are 2.1%and 3.3%higher than those derived from the conventional convolutional *** the information of urban form is essential to urban climate models,the results presented in this study can become a basis to refine the quantification of urban climate parameters,thereby introducing the urban heterogeneity to help understand the climate response of future urbanization.

关键词： convolutional autoencoder Self-Organizing Map(SOM) Gaussian Mixture Model(GMM) urban form clustering

来源：评论

学校读者我要写书评

暂无评论

Plant Disease Detection and Classification Using Hybrid Model Based on convolutional Auto Encoder and convolutional Neural Network

引用

CMC-COMPUTERS MATERIALS & CONTINUA 2025年第3期83卷 5219-5234页

作者： Kumar, Tajinder Kaur, Sarbjit Sharma, Purushottam Chhikara, Ankita Cheng, Xiaochun Lalar, Sachin Verma, Vikram Jai Parkash Mukand Lal Innovat Engn & Technol Inst Comp Sci & Engn Dept Yamunanagar 135133 India Govt PG Coll Dept Comp Sci Ambala 134003 India Galgotias Univ Sch Comp Sci & Engn Greater Noida 203201 India Kurukshetra Univ Dept Comp Sci & Applicat Kurukshetra 136118 India Swansea Univ Comp Sci Dept Bay Campus Fabian Way Swansea SA1 8EN Wales Gurugram Univ Dept Engn & Technol Gurugram 122003 India Panipat Inst Engn & Technol Dept CSE Panipat 132103 India

During its growth stage, the plant is exposed to various diseases. Detection and early detection of crop diseases is a major challenge in the horticulture industry. Crop infections can harm total crop yield and reduce farmers' income if not identified early. Today's approved method involves a professional plant pathologist to diagnose the disease by visual inspection of the afflicted plant leaves. This is an excellent use case for Community Assessment and Treatment Services (CATS) due to the lengthy manual disease diagnosis process and the accuracy of identification is directly proportional to the skills of pathologists. An alternative to conventional Machine Learning (ML) methods, which require manual identification of parameters for exact results, is to develop a prototype that can be classified without pre-processing. To automatically diagnose tomato leaf disease, this research proposes a hybrid model using the convolutional Auto-Encoders (CAE) network and the CNN-based deep learning architecture of DenseNet. To date, none of the modern systems described in this paper have a combined model based on DenseNet, CAE, and convolutional Neural Network (CNN) to diagnose the ailments of tomato leaves automatically. The models were trained on a dataset obtained from the Plant Village repository. The dataset consisted of 9920 tomato leaves, and the model-to-model accuracy ratio was 98.35%. Unlike other approaches discussed in this paper, this hybrid strategy requires fewer training components. Therefore, the training time to classify plant diseases with the trained algorithm, as well as the training time to automatically detect the ailments of tomato leaves, is significantly reduced.

关键词： Tomato leaf disease deep learning DenseNet-121 convolutional autoencoder convolutional neural 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：