Robust deep convolutional autoencoder To address this issue, we propose a deep convolutional embedded clus-tering algorithm in this paper. Dec 17, 2024 · Applying deep learning to unsupervised bearing fault diagnosis in complex industrial environments is challenging. In this tutorial, we will take a closer look at autoencoders (AE). It integrates robust loss terms, explicit noise modeling, and architectural innovations like spatio-temporal encoding to enhance its resilience and performance. This paper contributes a new type of model-based deep convolutional autoencoder that joins forces of state-of-the-art generative and CNN-based regression approaches for dense 3D face reconstruction via a deep integration of the two on an architectural level. To address these two limitations, we propose robust and explainable unsupervised autoencoder Dec 30, 2017 · We compare our proposed Robust Convolutional Autoencoder (RCAE) with the following state-of-the art methods for anomaly detection: Truncated SVD, which for zero-mean features is equivalent to PCA. With the unsupervised model, it becomes possible to extract effective and discriminative features from a huge unlabeled data set, which makes this approach widely applicable for the extraction of features May 8, 2025 · An autoencoder (AE) is an important model in the arsenal of deep learning models because it generates compact representations of unlabeled data, using unsupervised training (Goodfellow et al. , robust PCA. This repo offers an implementation based on Tensorflow. To achieve this, we propose a variant of the convolutional autoencoder (CAE) called SCDAC, which incorporates sparse convolutional embedding and a two-stage application Feb 1, 2023 · In 2018, Zong et al. To address the above problems, a multi-scale convolutional autoencoder with attention mechanism (MSCAE-AM) is developed. Sep 6, 2018 · In this paper, we propose a clustering approach embedded in a deep convolutional auto-encoder (DCAE). py contains Convolution Autoencoder network architecture while python files with _AE contains Autoencoder network architecture used for models in the paper Apr 7, 2022 · Time series data occurs widely, and outlier detection is a fundamental problem in data mining, which has numerous applications. , 2011). The primary goal of this study is to address these issues by developing a more robust deep clustering method. However, Deep Learning models are vulnerable to noise and redundant information encapsulated into the high-dimensional raw input images, leading to unstable and unreliable predictions. Dec 18, 2021 · The autoencoder based on a deep neural network is considered as one of the most robust unsupervised learning models of the last few decades. Nevertheless, most prevalent image augmentation recipes confine themselves to off-the-shelf linear Apr 1, 2022 · In this paper, a denoising temporal convolutional recurrent autoencoder (DTCRAE) is proposed to improve the performance of the temporal convolutional … This project is an implementation of a Deep Convolutional Denoising Autoencoder to denoise corrupted images. DCAEs have been effective in image processing as it fully utilizes the properties of convolutional neural networks. , earthquake vs. Initially, a progressive multi-stage learning framework based on convolutional autoencoders is employed. Then, a clustering oriented loss is directly built on embedded features to jointly perform feature re nement and cluster assignment. Feb 13, 2018 · Robust autoencoder is a model that combines Autoencoder and Robust PCA which can detect both noise and outliers. As an alternative, this article presents an unsupervised DL method for erratic-plus-Gaussian noise removal based on a robust deep convolutional autoencoder (RDCAE). What about the deep autoencoder, as a nonlinear generalization of PCA? This further motivates us to “reinvent” a factorization-based PCA as well as its nonlinear generalization. To achieve this, we propose a variant of the convolutional autoencoder (CAE) called SCDAC, which incorporates sparse convolutional embedding and a two-stage application Nov 1, 2024 · To address this issue, a robust multi-stage progressive autoencoder for hyperspectral anomaly detection (RMSAD) is proposed. RCAE is a class of deep convolutional autoencoders designed for unsupervised feature learning and robust anomaly detection in noisy, corrupted environments. This paper explores the integration of H-divergences into deep convolutional autoencoder-based clustering, aiming to provide a sophisticated approach that not only addresses the inherent limitations of previous metrics but also leverages the strengths of modern deep learning architectures to advance the field of data clustering. Most HSU methods adopt the 2D model for simplicity, whereas the performance of HSU depends on spectral and spatial information. org e-Print archive for research papers on various topics, including time-series forecasting and autoencoders. Abstract PCA can be made robust to data corruption, i. Denoising helps the autoencoders to learn the latent representation present in the data. This paper proposes a novel unsupervised approach, WDCAE-LKA, combining a wide kernel convolutional autoencoder (WDCAE) with a large kernel attention (LKA) mechanism to improve Sep 1, 2024 · To address the issues, we propose a Robust Tensor Convolutional Autoencoder (RTCAE) where the autoencoder instead of SVD is exploited to recover the normal data from the corrupted measurement tensor data. [32] presented a temporal convolutional network autoencoder based on dilated convolution for unsupervised anomaly detection in electrocardiogram of patients with cardiac arrhythmia. Nov 15, 2023 · To address the issues, we propose a Robust Tensor Convolutional Autoencoder (RTCAE) where the autoencoder instead of SVD is exploited to recover the normal data from the corrupted measurement tensor data. contains the code and datasets used for models in the paper Robust, Deep and Inductive Anomaly Detection The python files with _CAE. Aug 4, 2017 · Herein, we demonstrate novel extensions to deep autoencoders which not only maintain a deep autoencoders' ability to discover high quality, non-linear features but can also eliminate outliers and noise without access to any clean training data. - AyanS2004/Image-Steganography-Autoencoder Mar 6, 2024 · In addition, the limited labeled data makes the training of deep network models even more challenging. [33] constructed an unsupervised fake news detection method based on autoencoder for detecting anomalies on social networks. In this paper, we present a new method for robust supervised HSU based on a deep hybrid (3D and 2D) convolutional autoencoder (DHCAE) network. In this paper, we designed tests to evaluate this idea of using autoencoders as feature extractors for different seismological applications, such as event discrimination (i. For linear Jul 1, 2022 · In this work, the four convolutional layers of feature map at each time step are fed into ConvGRU autoencoder respectively, and then output the hidden pattern with the same dimension, prepared for decoding. explosion waveforms), and Oct 15, 2024 · Digital watermarking stands out as a pivotal solution for image security. Autoencoders are trained on encoding input data such as images into a smaller feature vector, and afterward, reconstruct it by a second neural network, called a decoder. A robust, production-ready image-in-image steganography system using deep convolutional autoencoders. Feb 13, 2025 · Existing deep clustering approaches often struggle with redundant feature learning, which limits their effectiveness. Fo-cusing on sparse corruption, we model the sparsity structure explicitly using the `1 norm to obtain var-ious robust formulations. overs the compression of underwater images based on a deep convolutional autoencoder. Our review focuses on the innovative deep watermarking approaches that employ neural networks to identify robust embedding spaces, resilient to various attacks. , 2016; Ng et al. The proposed image compression approach is studied using performance parameters 3 From robust PCA to robust autoencoders We now present our robust (convolutional) autoencoder model for anomaly de-tection. Mar 15, 2025 · This dual-training approach makes the autoencoder to be robust against potential manipulations or anomalies in data, thereby enhancing security and reliability, especially in sensitive applications like medical imaging and fraud detection. May 1, 2024 · Request PDF | On May 1, 2024, Aymane Bouali and others published Robust deep image clustering using convolutional autoencoder with separable discrete Krawtchouk and Hahn orthogonal moments | Find Sep 6, 2022 · In this paper, we present a new method for robust supervised HSU based on a deep hybrid (3D and 2D) convolutional autoencoder (DHCAE) network. The method can be seen as an extension of robust PCA to allow for a nonlinear manifold that explains most of the data. In 2021, Li et al. Denoising autoencoders ensures a good representation is one This paper explores the integration of H-divergences into deep convolutional autoencoder-based clustering, aiming to provide a sophisticated approach that not only addresses the inherent limitations of previous metrics but also leverages the strengths of modern deep learning architectures to advance the field of data clustering. Oct 15, 2024 · OmniAnomaly could provide an intuitive and effective method to interpret the detected entity anomalies based on the reconstruction probability. The con ept of underwater image acquisition techniques and their analysis are also discussed. Specifically, as a typical unsupervised learning model, the encoder can effectively reduce the dependence on labeled data. [30] proposed a deep autoencoding Gaussian mixture model (DAGMM) by fusing autoencoder and Gaussian mixture model that uses the joint optimization form to solve the problem of decoupling model learning and achieves good unsupervised anomaly detection results on public benchmark datasets. Autoencoders constitute an unsupervised dimensionality reduction technique, proven to filter out noise and 3 From robust PCA to robust autoencoders We now present our robust (convolutional) autoencoder model for anomaly de-tection. Supports multiple model variants, advanced loss functions, robustness testing, and web deployment via Flask. Speci cally, we develop a convolutional autoencoders structure to learn embedded features in an end-to-end way. Our method consists of Feb 1, 2023 · In 2018, Zong et al. To this end, regularization techniques like image augmentation are necessary for deep neural networks to generalize well. Jun 10, 2022 · Deep neural networks are capable of learning powerful representations to tackle complex vision tasks but expose undesirable properties like the over-fitting issue. . In contrast to conventional clustering approaches, our method simultaneously learns feature representations and cluster assignments through DCAEs. With the advent of deep learning, watermarking has seen significant advancements. In this work, we propose an ML topology for performing efficient and robust classification in high-dimensional and noisy input data images. Jun 1, 2024 · Creating a robust convolutional model “RDEICSKHM” using discrete separable moments. e. Traditional fault detection methods rely on labeled data, which is costly and labor-intensive to obtain. Zavrak et al. The noise level is not needed to be known. Dec 1, 2021 · The idea of using a deep autoencoder to encode seismic waveform features and then use them in different seismological applications is appealing. A novel reduced deep convolutional neural network (RDCNN) embedded with stack autoencoder, that is, RDCSAE structure is introduced to extract the most discriminative unsupervised feature data by importing the selected BLIMF of parameter-adaptive VMD (PAVMD) algorithm. As an alternative, following the RPCA framework, this article proposes an unsupervised iterative robust deep convolutional autoencoder (IRDCAE) model to suppress intense VSP coupling noise without any assumptions regarding valuable signals. Nov 20, 2021 · Deep convolutional neural networks have shown remarkable performance in the image classification domain. By cooperatively learning features and assigning clusters, deep clustering is superior to conventional clustering algorithms. While pruning has been extensively studied in convolutional neural networks (CNNs) and classification tasks, its application to AEs and clustering tasks remains relatively Explore the arXiv. noise waveforms, earthquake vs. Feb 1, 2023 · In 2021, Thill et al. Existing autoencoder-based approaches deliver state-of-the-art performance on challenging real-world data but are vulnerable to outliers and exhibit low explainability. [22] analyzed the detection capabilities of autoencoder and variational autoencoder deep learning methods alongside the one-class SVM using a semi-supervised strategy. iopf cz8hmy odbbsw xp6f mjcp6r ez5e 60cau rmjwoc 5d7surm cf8ka8