检索结果-内蒙古大学图书馆

Binarizing historical architectural drawings with shallow convolutional autoencoders

ENGINEERING APPLICATIONS OF ARTIFICIAL INTELLIGENCE 2025年 148卷

作者： Narag, Mark Jeremy G. Lico, Gerard Rey Soriano, Maricor Univ Philippines Diliman Natl Inst Phys Quezon City 1101 Philippines Univ Philippines Diliman Coll Architecture Quezon City 1101 Philippines

We developed convolutional autoencoder models to binarize 26 severely stained 1938 architectural drawings of the Department of Agriculture and Commerce Building in Manila. Due to the presence of deep-seated stains that have accumulated over time on the original linen drawings, traditional thresholding techniques failed in binarizing them. We propose utilizing the capabilities of deep learning models, particularly convolutional autoencoders, to automate the binarization of all 26 drawings. Our goal is to efficiently train a model using a segment from a single drawing, thereby eliminating the requirement for separate training on each individual drawing. We manually binarized a section of one drawing to serve as the desired output within our training dataset. Moreover, since deep learning models are computationally heavy, we proposed simplifying them into shallow models by trimming down the architecture to a single hidden layer. We created four pairs of models (deep and shallow) with input sizes 32 x 32, 64 x 64, 128 x 128, and 256 x 256. All the models effectively binarized the drawing by successfully separating the pronounced stains from the drawn lines, which the traditional binarization techniques have failed to do. The models achieved F1 scores from 0.961 to 0.977, intersection-over-union scores from 0.925 to 0.955, and peak-signal-to-noise ratio values from 23.1 to 25.4. Notably, we achieved higher performance metrics on shallow autoencoders than their deep counterparts. Our workflow also succeeded in binarizing a separate collection of line art drawings on vellum by another architect, which were at a lower resolution.

关键词： Binarization convolutional autoencoder Architectural drawing Line art

来源：评论

学校读者我要写书评

暂无评论

Multi-resolution reconstruction of longitudinal streambed footprints using embedded sparse convolutional autoencoders

引用

JOURNAL OF HYDROLOGY 2025年 654卷

作者： Yang, Yifan Tang, Zihao Shao, Dong Xu, Zhonghou Wuhan Univ State Key Lab Water Resources Engn & Management Wuhan Peoples R China Univ Auckland Dept Civil & Environm Engn Auckland New Zealand Chinese Acad Sci Inst Mech Key Lab Mech Fluid Solid Coupling Syst Beijing Peoples R China Natl Inst Water & Atmospher Res Hamilton New Zealand

This study introduces an embedded convolutional autoencoder (CAE) architecture designed for the multiresolution reconstruction of longitudinal streambed footprints as sparse heatmaps. Three standalone but interrelated CAEs are trained to achieve double-upsampling, enhancing the heatmaps' spatial resolution and data measurement resolution simultaneously. Transfer learning improves model training efficiency by incorporating a trained model into a larger model at the next level. Cascading the CAEs facilitates a direct pathway for enhancing data quality from the coarsest inputs and recovering fine-grained patterns. Systematic evaluations prove the CAEs' reliability in working individually and collectively. Robustness analyses demonstrate the model's ability to retain field reconstructive quality when subjected to various corrupted inputs, including bulk data loss and spiky noise interference with local measurements at different streambed sections. The model's capacity benefitted from including attention mechanisms (convolutional block attention modules, CBAM) and the adaptive training strategy using crafted loss functions, ensuring efficient extraction and learning of sparse dense patterns and fast reconstruction of physically sound fields. The model architecture's flexibility and scalability are highlighted, proving it suitable for more complex geophysical systems with higher dimensions. The proposed embedded CAE architecture provides a foundational tool for creating digital surrogates of river courses and similar entities, which often involve inherently sparsely distributive data in both spatial and temporal domains.

关键词： Machine learning Streambed footprint convolutional autoencoder Multi-resolution reconstruction

来源：评论

学校读者我要写书评

暂无评论

Anomaly Detection in Batch Manufacturing Processes Using Localized Reconstruction Errors From 1-D convolutional autoencoders

引用

IEEE TRANSACTIONS ON SEMICONDUCTOR MANUFACTURING 2023年第1期36卷 147-150页

作者： Gorman, Mark Ding, Xuemei Maguire, Liam Coyle, Damien Intelligent Syst Res Ctr Data Sci Grp Sch Comp Engn & Intelligent Syst Londonderry BT48 0LY North Ireland Ulster Univ Intelligent Syst Res Ctr Sch Comp Engn & Intelligent Syst Londonderry BT48 7JL North Ireland

Multivariate batch time-series data sets within Semiconductor manufacturing processes present a difficult environment for effective Anomaly Detection (AD). The challenge is amplified by the limited availability of ground truth labelled data. In scenarios where AD is possible, black box modelling approaches constrain model interpretability. These challenges obstruct the widespread adoption of Deep Learning solutions. The objective of the study is to demonstrate an AD approach which employs 1-Dimensional convolutional autoencoders (1d-CAE) and Localised Reconstruction Error (LRE) to improve AD performance and interpretability. Using LRE to identify sensors and data that result in the anomaly, the explainability of the Deep Learning solution is enhanced. The Tennessee Eastman Process (TEP) and LAM 9600 Metal Etcher datasets have been utilised to validate the proposed framework. The results show that the proposed LRE approach outperforms global reconstruction errors for similar model architectures achieving an AUC of 1.00. The proposed unsupervised learning approach with AE and LRE improves model explainability which is expected to be beneficial for deployment in semiconductor manufacturing where interpretable and trustworthy results are critical for process engineering teams.

关键词： Feature extraction Sensors Training Semiconductor device modeling Deep learning Shape Metals fault detection and classification semiconductor manufacturing convolutional autoencoder reconstruction error

来源：评论

学校读者我要写书评

暂无评论

Basic research on usefulness of convolutional autoencoders in detecting defects in concrete using hammering sound

引用

STRUCTURAL HEALTH MONITORING-AN INTERNATIONAL JOURNAL 2023年第4期22卷 2231-2250页

作者： Sonoda, Yoshimi Lu, Chi Yin, Yifan Kyushu Univ Fukuoka Japan

Because hammering sound tests are inexpensive and can be performed easily, they are commonly used as an inspection method for examining the presence of defect areas (voids or peelings) in aged concrete structures. However, the evaluation of the health of concrete using hammering sounds depends on the subjective experience of the inspector. Therefore, there is a demand to develop a highly reliable and objective diagnostic method that is accurate and efficient. In this study, we used a convolutional autoencoder (CAE) to develop a diagnostic method that could assist the inspectors with quantitative diagnostic results of tapping sound when detecting defect areas in concrete. In particular, we verified the anomaly detection accuracy of hammering sound data of actual bridges that have deteriorated over time using the proposed CAE model.

关键词： Hammering sound test concrete flaking anomaly detection convolutional autoencoder time-frequency analysis

来源：评论

学校读者我要写书评

暂无评论

convolutional autoencoders and CGANs for unsupervised structural damage localization

引用

MECHANICAL SYSTEMS AND SIGNAL PROCESSING 2024年 220卷

作者： Junges, Rafael Rastin, Zahra Lomazzi, Luca Giglio, Marco Cadini, Francesco Politecn Milan Dipartimento Meccan Via La Masa 1 I-20156 Milan MI Italy

The present work introduces two unsupervised data -driven methodologies for processing Lamb waves (LWs) to localize structural damage, specifically employing convolutional autoencoders (CAEs) and conditional generative adversarial networks (CGANs). Both techniques are capable of processing diagnostic signals without the need for any prior feature extraction. Once all signals are processed, a damage probability map is generated. The performance of the methods was tested using two different experimental datasets. The first derives from LWs obtained from a set of piezoelectric transducers mounted on two different composite panels, made of two different layups. Pseudo -damage and real damage were considered. The second dataset derives from LWs acquired on a full-scale composite wing, where damage was introduced through impacts performed using an air -gun. The results of this study revealed that the proposed unsupervised methods are capable of localizing damage properly, with comparable accuracy.

关键词： Unsupervised deep learning convolutional autoencoder Generative adversarial network Damage localization Lamb waves

来源：评论

学校读者我要写书评

暂无评论

convolutional block attention autoencoder for denoising electrocardiograms

引用

BIOMEDICAL SIGNAL PROCESSING AND CONTROL 2023年第PartB期86卷

作者： Chorney, Wesley Wang, Haifeng He, Lu Lee, Seunghan Fan, Lir-Wan Mississippi State Univ Computat Engn Mississippi State MS 39762 USA Mississippi State Univ Ind & Syst Engn Mississippi State MS 39762 USA Mississippie State Univ Mkt Quantitat Anal & Business Law Mississippie State MS 39762 USA Univ Mississippi Pediat Newborn Med Med Ctr Jackson MS 39216 USA

Electrocardiograms are commonly used to detect cardiovascular diseases, so it is important that they are of high quality. However, various sources of noise, such as baseline wander, muscle artifact, and electrode motion can obfuscate the signal of the heart recorded by different monitors. In this work, we propose a novel algorithm, convolutional denoising autoencoder with block attention module (CDAE-BAM), for removing noise from electrocardiograms by leveraging attention in a convolutional denoising autoencoder. We propose an attention block including both spatial and channel attention. Spatial attention captures the location of relevant features within channels in a signal, and channel attention captures the most relevant channels in a signal. We validate our proposed algorithm's performance in removing noise from the MIT-BIH Noise Stress Test Database from electrocardiogram signals in the QT Database, the Computing in Cardiology Challenge 2017 Database, and the Medical Information Mart for Intensive Care Database. We show that this method outperforms eight other state-of-the-art methods with respect to sum of squared distances, mean absolute distance, and cosine similarity.

关键词： Denoising Electrocardiograms convolutional autoencoder Attention

来源：评论

学校读者我要写书评

暂无评论

Unsupervised convolutional autoencoders for 4D transperineal ultrasound classification

引用

JOURNAL OF MEDICAL IMAGING 2023年第1期10卷 014004页

作者： van den Noort, Frieda Manzini, Claudia Hofsteenge, Merijn Sirmacek, Beril van der Vaart, Carl H. H. Slump, Cornelis H. H. Univ Twente Tech Med Ctr Robot & Mechatron Fac Elect Engn Math & Comp Sci Enschede Netherlands Univ Med Ctr Utrecht Dept Obstet & Gynecol Utrecht Netherlands Saxion Univ Appl Sci Sch Creat Technol Smart Cities Grp Enschede Netherlands

Purpose: 4D Transperineal ultrasound (TPUS) is used to examine female pelvic floor disorders. Muscle movement, like performing a muscle contraction or a Valsalva maneuver, can be captured on TPUS. Our work investigates the possibility for unsupervised analysis and classification of the TPUS ***: An unsupervised 3D-convolutional autoencoder is trained to compress TPUS volume frames into a latent feature vector (LFV) of 128 elements. The (co)variance of the features are analyzed and statistical tests are performed to analyze how features contribute in storing contraction and Valsalva information. Further dimensionality reduction is applied (principal component analysis or a 2D-convolutional autoencoder) to the LFVs of the frames of the TPUS movie to compress the data and analyze the interframe movement. Clustering algorithms (K-means clustering and Gaussian mixture models) are applied to this representation of the data to investigate the possibilities of unsupervised ***: The majority of the features show a significant difference between contraction and Valsalva. The (co)variance of the features from the LFVs was investigated and features most prominent in capturing muscle movement were identified. Furthermore, the first principal component of the frames from a single TPUS movie can be used to identify movement between the frames. The best classification results were obtained after applying principal component analysis and Gaussian mixture models to the LFVs of the TPUS movies, yielding a 91.2% ***: Unsupervised analysis and classification of TPUS data yields relevant information about the type and amount of muscle movement present.

关键词： urogynecology transperineal ultrasound convolutional autoencoder classification unsupervised learning

来源：评论

学校读者我要写书评

暂无评论

Early detection of internal erosion in earth dams: combining seismic monitoring and convolutional autoencoders

引用

GEORISK-ASSESSMENT AND MANAGEMENT OF RISK FOR ENGINEERED SYSTEMS AND GEOHAZARDS 2023年第3期18卷 570-590页

作者： Yousefpour, Negin Fazel Mojtahedi, Farid Univ Melbourne Dept Infrastruct Engn Parkville Vic Australia

Levees/earth dams are critical infrastructures for supplementing clean water, flood management, and energy production, prone to progressive failures due to internal erosion. Current inspection methods are unable to detect internal erosion until its exterior manifestation when it is too late to prevent the often-catastrophic failures. Therefore, finding innovative methods for the early detection of internal erosion is crucial. Despite the knowledge about the general mechanism of internal erosion, its early detection (and prevention) has remained a gap. This study introduces a novel artificial intelligence (AI) method to identify the temporal patterns within the passive seismic monitoring data, which can be associated with internal erosion initiation in earth dams. The proposed approach implements convolutional autoencoders, an emerging deep-learning algorithm for anomaly detection in time-series data. Through an unsupervised learning framework, the autoencoders are trained using passive seismic monitoring data collected from a full-scale test embankment. In addition to the approximate initiation time, this algorithm can evaluate the initiation location by identifying the first sensors demonstrating internal erosion signs. The proposed deep learning framework combined with continuous seismic monitoring can serve as a basis for developing advanced early warning systems for internal erosion in earth dams.

关键词： Internal erosion Passive seismic data Anomaly detection convolutional autoencoder Artificial intelligence Earth dams

来源：评论

学校读者我要写书评

暂无评论

Anomaly detection of sensor faults and extreme events by anomaly locating strategies and convolutional autoencoders

引用

STRUCTURAL HEALTH MONITORING-AN INTERNATIONAL JOURNAL 2025年

作者： Zhang, Yuxuan Wang, Xiaoyou Wu, Wanglin Xia, Yong Hong Kong Polytech Univ Dept Civil & Environm Engn ZS9639-FBlock Z Hong Kong Peoples R China Hong Kong Polytech Univ Joint Res Ctr Marine Infrastructure Hong Kong Peoples R China Minist Emergency Management China Sichuan Fire Res Inst Chengdu Peoples R China

Deep learning has demonstrated advantages in the detection and classification of sensor faults. However, challenges exist in the domain of extreme event detection, including the scarcity of training data due to the rarity of extreme events and the difficulty of distinguishing between sensor anomalies and extreme events. This study proposes a novel two-stage anomaly detection method that can separately detect extreme events and sensor faults from the normal pattern without using any data related to extreme events in the training step. In the first stage, the potential segments are located using three defined indexes from raw normal acceleration data and then represented as peak and root mean-square envelopes. Then, these envelopes are normalized and grouped by the three indexes to train three convolutional autoencoders. The preprocessing strategies minimize the difference of the normalized sequence shapes between the normal pattern and extreme response. Hence, well-trained convolutional autoencoders can detect sensor fault patterns with different sequence shapes from the normal pattern. In the second stage, several thresholds are defined to separate the patterns sensitive only to absolute values from the normal pattern, including extreme events and partial Minor scenarios. A multilabel classification is followed to identify specific sensor fault patterns. Datasets from two real structural health monitoring systems are utilized to validate the proposed method. Results show that extreme events and sensor faults can be detected efficiently and accurately. Three extreme events, Typhoon Saola, Typhoon Higos, and a pedestrian test, are successfully detected using the developed method.

关键词： Structural health monitoring data anomaly detection extreme event sensor fault convolutional autoencoder

来源：评论

学校读者我要写书评

暂无评论

Compression of EMG Signals Using Deep convolutional autoencoders

引用

IEEE JOURNAL OF BIOMEDICAL AND HEALTH INFORMATICS 2022年第7期26卷 2888-2897页

作者： Dinashi, Kimia Ameri, Ali Akhaee, Mohammad Ali Englehart, Kevin Scheme, Erik Univ Tehran Dept Elect Engn Tehran 111554563 Iran Shahid Beheshti Univ Med Sci Sch Med Dept Biomed Engn Tehran *** Iran Univ New Brunswick Inst Biomed Engn Fredericton NB E3B 5A3 Canada

Efficient storage and transmission of electromyogram (EMG) data are important for emerging applications such as telemedicine and big data, as a vital tool for further advancement of the field. However, due to limitations in internet speed and hardware resources, transmission and storage of EMG data are challenging. As a solution, this work proposes a new method for EMG data compression using deep convolutional autoencoders (CAE). Eight-channel EMG data from 10 subjects, and high-density EMG data from 18 subjects, were investigated for compression. The CAE architecture was designed to extract an abstract data representation that is heavily compressed, but from which the salient information for classification can be effectively reconstructed. The proposed method attained efficient compression;for CR = 1600, the average PRDN (percentage RMS difference normalized) was 31.5% and the wrist motions classification accuracy (CA) reduced roughly 5%. The CAE substantially outperformed the state-of-the-art high-efficiency video coding and a well-known wavelet-thresholding compression technique. Moreover, by reducing the bit-resolution of the CAE's compressed data from 24 bits to 6 bits, an additional 4-fold compression was achieved without significant degradation of the reconstruction performance. Furthermore, the CAE's inter-subject performance was promising;e.g., for CR = 1600, the PRDN for the inter-subject case was only 2.6% less than that of the within-subject performance. The powerful EMG compression performance with remarkable reconstruction results reflects the CAEs potential as an automatic end-to-end approach with the ability to learn the complete encoding and decoding process. Furthermore, the excellent inter-subject performance demonstrates the generalizability and usability of the proposed approach.

关键词： Electromyography convolutional codes Convolution Training Encoding Electrodes Wrist Big data compression convolutional autoencoder deep learning EMG

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

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

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：