Convolutional Autoencoder Research Articles

Optimizing dotted Arabic expiration date recognition with ARABEX: a convolutional autoencoder with bidirectional LSTM and CRNN approach

Optimizing dotted Arabic expiration date recognition with ARABEX: a convolutional autoencoder with bidirectional LSTM and CRNN approach

Whole slide image based prognosis prediction in rectal cancer using unsupervised artificial intelligence

BackgroundRectal cancer is a common cancer worldwide and lacks effective prognostic markers. The development of prognostic markers by computational pathology methods has attracted increasing attention. This paper aims to construct a prognostic signature from whole slide images for predicting progression-free survival (PFS) of rectal cancer through an unsupervised artificial intelligence algorithm.MethodsA total of 238 patients with rectal cancer from two datasets were collected for the development and validation of the prognostic signature. A tumor detection model was built by transfer learning. Then, on the basis of the tumor patches recognized by the tumor detection model, a convolutional autoencoder model was built for decoding the tumor patches into deep latent features. Next, on the basis of the deep latent features, the tumor patches were divided into different clusters. The cluster number and other hyperparameters were optimized by a nested cross-validation method. The percentage of each cluster from the patient’s tumor patches, which is hereafter called PCF, was calculated for prognostic signature construction. The prognostic signature was constructed by Cox proportional hazard regression with L2 regularization. Finally, bioinformatic analysis was performed to explore the underlying biological mechanisms of the PCFs.ResultsThe accuracy of the tumor detection model in distinguishing tumor patches from non-tumor patches achieved 99.3%. The optimal cluster number was determined to be 9. Therfore, 9 PCFs were calculated to construct the prognostic signature. The prognostic signature achieved a concordance index of 0.701 in the validation cohort. The Kaplan-Meier survival curves showed the prognostic signature had good risk stratification ability. Through the bioinformatic analysis, several PCF-associated genes were identified. These genes were enriched in various gene ontology terms.ConclusionThe developed prognostic signature can effectively predict PFS in patients with rectal cancer and exploration of the underlying biological mechanisms may help to promote its clinical translation.

Comparative Analysis of Denoising Methods to Improve Image Quality for Medical Visual Question Answering

Medical Visual Question Answering (MedVQA) is a dual research field that intersects medical imaging and natural language processing, for better interpretability and accessibility of medical image data.Medical image quality is paramount for accurate diagnostics and subsequent medical visual question answering (MedVQA) tasks. This research focuses on applying and then analyzing results of different denoising methods on VQA-RAD, Medical VQA dataset to enhance quality of images. This study explores effectiveness of different traditional and deep learning based methods to reduce noise within medical images, thereby improving the accuracy and reliability of MedVQA task. We applied different traditional denoising filtering methods such as, gaussian filter, median filter, average filter, bilateral filter and convolutional autoencoder (CAE) based on deep learning on a VQA-RAD dataset to compare effectiveness of each denoising methods to improve image quality. Through comprehensive experiments and evaluations, this paper demonstrates that the convolutional autoencoder is potentially enhancing quality of medical images with an emphasis on preserving essential diagnostic information while suppressing unwanted noise with compare to other traditional denoising filters. The denoised images are then employed as input to improve accuracy for MedVQA tasks. The results of this research will help in optimizing medical imaging pipelines, ultimately benefiting clinical decision-making and healthcare outcomes.

Unsupervised Learning for Machinery Adaptive Fault Detection Using Wide-Deep Convolutional Autoencoder with Kernelized Attention Mechanism

Applying deep learning to unsupervised bearing fault diagnosis in complex industrial environments is challenging. Traditional fault detection methods rely on labeled data, which is costly and labor-intensive to obtain. 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 fault detection under unlabeled conditions, and the adaptive threshold module based on a multi-layer perceptron (MLP) dynamically adjusts thresholds, boosting model robustness in imbalanced scenarios. Experimental validation on two datasets (CWRU and a customized ball screw dataset) demonstrates that the proposed model outperforms both traditional and state-of-the-art methods. Notably, WDCAE-LKA achieved an average diagnostic accuracy of 90.29% in varying fault scenarios on the CWRU dataset and 72.89% in the customized ball screw dataset and showed remarkable robustness under imbalanced conditions; compared with advanced models, it shortens training time by 10–26% and improves average fault diagnosis accuracy by 5–10%. The results underscore the potential of the WDCAE-LKA model as a robust and effective solution for intelligent fault diagnosis in industrial applications.

An efficient unsupervised classification model for galaxy morphology: Voting clustering based on coding from ConvNeXt large model

By combining unsupervised and supervised machine learning methods, we have proposed a framework, called USmorph to carry out automatic classifications of galaxy morphologies. In this work, we update the unsupervised machine learning (UML) step by proposing an algorithm based on ConvNeXt large model coding to improve the efficiency of unlabeled galaxy morphology classifications. The method can be summarized into three key aspects as follows: (1) a convolutional autoencoder is used for image denoising and reconstruction and the rotational invariance of the model is improved by polar coordinate extension; (2) utilizing a pre-trained convolutional neural network (CNN) named ConvNeXt for encoding the image data. The features were further compressed via a principal component analysis (PCA) dimensionality reduction; (3) adopting a bagging-based multi-model voting classification algorithm to enhance robustness. We applied this model to I-band images of a galaxy sample with $I_ mag < 25$ in the COSMOS field. Compared to the original unsupervised method, the number of clustering groups required by the new method is reduced from 100 to 20. Finally, we managed to classify about 53<!PCT!> galaxies, significantly improving the classification efficiency. To verify the validity of the morphological classification, we selected massive galaxies with $M_ M_ sun $ for morphological parameter tests. The corresponding rules between the classification results and the physical properties of galaxies on multiple parameter surfaces are consistent with the existing evolution model. Our method has demonstrated the feasibility of using large model encoding to classify galaxy morphology, which not only improves the efficiency of galaxy morphology classification, but also saves time and manpower. Furthermore, in comparison to the original UML model, the enhanced classification performance is more evident in qualitative analysis and has successfully surpassed a greater number of parameter tests. The enhanced UML method will support the Chinese space station telescope in the future.

Ensemble reconstruction of missing satellite data using a denoising diffusion model: application to chlorophyll a concentration in the Black Sea

Abstract. Satellite observations provide a global or near-global coverage of the World Ocean. They are however affected by clouds (among others), which severely reduce their spatial coverage. Different methods have been proposed in the literature to reconstruct missing data in satellite observations. For many applications of satellite observations, it has been increasingly important to accurately reflect the underlying uncertainty of the reconstructed observations. In this paper, we investigate the use of a denoising diffusion model to reconstruct missing observations. Such methods can naturally provide an ensemble of reconstructions where each member is spatially coherent with the scales of variability and with the available data. Rather than providing a single reconstruction, an ensemble of possible reconstructions can be computed, and the ensemble spread reflects the underlying uncertainty. We show how this method can be trained from a collection of satellite data without requiring a prior interpolation of missing data and without resorting to data from a numerical model. The reconstruction method is tested with chlorophyll a concentration from the Ocean and Land Colour Instrument (OLCI) sensor (aboard the satellites Sentinel-3A and Sentinel-3B) on a small area of the Black Sea and compared with the neural network DINCAE (Data-INterpolating Convolutional Auto-Encoder). The spatial scales of the reconstructed data are assessed via a variogram, and the accuracy and statistical validity of the reconstructed ensemble are quantified using the continuous ranked probability score and its decomposition into reliability, resolution, and uncertainty.

Machine Learning Based Prediction of Ditching Loads

Approaches are presented to predict dynamic ditching loads on aircraft fuselages using machine learning. The employed learning procedure is structured into two parts, the reconstruction of the spatial loads using a convolutional autoencoder (CAE) and the transient evolution of these loads in a subsequent part. Different CAE strategies are assessed and combined with either long short-term memory (LSTM) networks or Koopman operator based methods to predict the transient behavior. The training data are compiled by an extension of the momentum method of von Karman and Wagner, and the rationale of the training approach is briefly summarized. The application included refers to a full-scale fuselage of a DLR-D150 aircraft for a range of horizontal and vertical approach velocities at 6 deg incidence. Results indicate a satisfactory level of predictive agreement for all four investigated surrogate models examined, with the combination of an LSTM and a deep decoder CAE showing the best performance.

From data to dynamics: Reconstructing soliton collision phenomena in optical fibers using a convolutional autoencoder

In this study, a convolutional autoencoder is constructed to extract and reconstruct the dynamical processes of soliton collisions in optical fibers. The model demonstrates exceptional reconstruction capabilities, accurately capturing the evolution of optical event horizons and reproducing nonlinear phenomena such as complex frequency conversions and energy exchange processes. The reconstruction results show high consistency with the numerical simulations, with RMSE values of 0.0220 and 0.0174 in the temporal and frequency domains, respectively. Additionally, by adjusting the training parameters of the convolutional autoencoder model, its reconstruction performance for nonlinear dynamic processes was further validated. This method is expected to provide a different perspective for studying nonlinear phenomena in optical fibers while reducing the consumption of computational resources.

Advanced image reconstruction for electrostatic tomography in gas-solid two-phase flow based on convolutional autoencoder neural network

The image reconstruction of flowing charged particles in gas-solid two-phase (GSTP) flow can be achieved through electrostatic tomography (EST). Accurate image reconstruction is crucial for detecting the movement patterns of the particles. In order to improve the quality of reconstructed images, a unique convolutional autoencoder neural network (CANN) is proposed. This study uses an image set generated by the linear backprojection (LBP) algorithm to train the CANN, which consists of an encoder and a decoder. The encoder utilizes convolutional and max-pooling layers to reduce the dimensionality of the images and extract key features, while the decoder restores and reconstructs the images through up-sampling and convolutional operations to closely approximate the reference image. To prevent overfitting, dropout layers are introduced after each max-pooling layer in the encoder. To verify the anti-noise capability of the network, Gaussian white noise ranging from 10 dB to 20 dB is added to the test set. The proposed CANN has been validated through simulations and experiments, demonstrating its effectiveness in overcoming noticeable artifacts and noise in reconstructed images when identifying GSTP flow patterns. Furthermore, it shows significant enhancements in imaging outcomes compared to conventional image reconstruction techniques and some current deep learning algorithms.

Data-driven reduced order surrogate modeling for coronary in-stent restenosis

Background:The intricate process of coronary in-stent restenosis (ISR) involves the interplay between different mediators, including platelet-derived growth factor, transforming growth factor-β, extracellular matrix, smooth muscle cells, endothelial cells, and drug elution from the stent. Modeling such complex multiphysics phenomena demands extensive computational resources and time. Methods:This paper proposes a novel non-intrusive data-driven reduced order modeling approach for the underlying multiphysics time-dependent parametrized problem. In the offline phase, a 3D convolutional autoencoder, comprising an encoder and decoder, is trained to achieve dimensionality reduction. The encoder condenses the full-order solution into a lower-dimensional latent space, while the decoder facilitates the reconstruction of the full solution from the latent space. To deal with the 5D input datasets (3D geometry + time series + multiple output channels), two ingredients are explored. The first approach incorporates time as an additional parameter and applies 3D convolution on individual time steps, encoding a distinct latent variable for each parameter instance within each time step. The second approach reshapes the 3D geometry into a 2D plane along a less interactive axis and stacks all time steps in the third direction for each parameter instance. This rearrangement generates a larger and complete dataset for one parameter instance, resulting in a singular latent variable across the entire discrete time-series. In both approaches, the multiple outputs are considered automatically in the convolutions. Moreover, Gaussian process regression is applied to establish correlations between the latent variable and the input parameter. Results:The constitutive model reveals a significant acceleration in neointimal growth between 30−60 days post percutaneous coronary intervention (PCI). The surrogate models applying both approaches exhibit high accuracy in pointwise error, with the first approach showcasing smaller errors across the entire evaluation period for all outputs. The parameter study on drug dosage against ISR rates provides noteworthy insights of neointimal growth, where the nonlinear dependence of ISR rates on the peak drug flux exhibits intriguing periodic patterns. Applying the trained model, the rate of ISR is effectively evaluated, and the optimal parameter range for drug dosage is identified. Conclusion:The demonstrated non-intrusive reduced order surrogate model proves to be a powerful tool for predicting ISR outcomes. Moreover, the proposed method lays the foundation for real-time simulations and optimization of PCI parameters.

Enhancing SNR in CEST imaging: A deep learning approach with a denoising convolutional autoencoder.

To develop a SNR enhancement method for CEST imaging using a denoising convolutional autoencoder (DCAE) and compare its performance with state-of-the-art denoising methods. The DCAE-CEST model encompasses an encoder and a decoder network. The encoder learns features from the input CEST Z-spectrum via a series of one-dimensional convolutions, nonlinearity applications, and pooling. Subsequently, the decoder reconstructs an output denoised Z-spectrum using a series of up-sampling and convolution layers. The DCAE-CEST model underwent multistage training in an environment constrained by Kullback-Leibler divergence, while ensuring data adaptability through context learning using Principal Component Analysis-processed Z-spectrum as a reference. The model was trained using simulated Z-spectra, and its performance was evaluated using both simulated data and in vivo data from an animal tumor model. Maps of amide proton transfer (APT) and nuclear Overhauser enhancement (NOE) effects were quantified using the multiple-pool Lorentzian fit, along with an apparent exchange-dependent relaxation metric. In digital phantom experiments, the DCAE-CEST method exhibited superior performance, surpassing existing denoising techniques, as indicated by the peak SNR and Structural Similarity Index. Additionally, in vivo data further confirm the effectiveness of the DCAE-CEST in denoising the APT and NOE maps when compared with other methods. Although no significant difference was observed in APT between tumors and normal tissues, there was a significant difference in NOE, consistent with previous findings. The DCAE-CEST can learn the most important features of the CEST Z-spectrum and provide the most effective denoising solution compared with other methods.

Temporal denoising and deep feature learning for enhanced defect detection in thermography using stacked denoising convolution autoencoder

Thermal wave imaging uses the temporal temperature distribution over the object’s surface for subsurface analysis. However, the noise generated during experimentation corrupts this temporal history and hampers the detection of defect signatures. As denoising of the temporal thermal history enhances the defect detectability, this study offers a stacked denoising convolution autoencoder (SDCAE) in frequency-modulated thermal wave imaging with one-dimensional convolution layers to reduce noise in temporal thermal evolution and train high-level features resulting in improved defect signs. Experimental results on mild steel and carbon fiber reinforced polymer specimens with different sizes of defects at various depths demonstrate that integrating temporal denoising and deep feature learning techniques into a single novel framework significantly improved defect detectability. In addition, defect signal-to-noise ratios of the denoised thermal data and latent space of the proposed model compared to conventional autoencoder and dimensionality reduction techniques recommend the superiority of the proposed method.

Diesel generator set anomalous sound detection based on lightweight convolutional autoencoder and prior knowledge integration

Diesel generator set anomalous sound detection based on lightweight convolutional autoencoder and prior knowledge integration

A digital twin framework for anomaly detection in industrial robot system based on multiple physics-informed hybrid convolutional autoencoder

The robot number in industry is growing up rapidly. Building anomaly detection system for them can improve the security of these expensive devices. The article implements an anomaly detection framework based on digital twin, which are built by a hybrid convolutional autoencoder. The framework shares those neural network weight files as digital assets, users can use them to estimate the possible output from real input. It approximates the dynamic relationship between motion, current, temperature and vibration with hybrid convolution. Considering the limited generalization performance of direct data-driven methods in practical physical systems, this article introduces physical information methods to improve the constraint function of neural network. The influence of multiple physical fields on current is established by a unified neural network. Terminals detect anomaly with KL divergence between really current and estimated current. The article collects operational data from real robots and verifies it, and the experiment shows that the RMSE for current estimation is below 1.5 %, the F1-score in anomaly detection is over 98.23 %, false positive is below 1 %, false negative is below 1.7 %. The relevant technologies are gradually being promoted and applied in enterprises.

Image-based flashback detection in a hydrogen-fired gas turbine using a convolutional autoencoder

Image-based flashback detection in a hydrogen-fired gas turbine using a convolutional autoencoder

Remaining useful life prediction of rolling bearings based on parallel feature extraction

Purpose This study aims to introduce a method for predicting the remaining useful life (RUL) of bearings based on parallel feature extraction. The proposed model provides prior knowledge and removes redundant handcrafted feature information, additionally, which focuses on the important features at different time scales. Design/methodology/approach Distinct from traditional parallel feature extraction methods, which can lead to information redundancy, a one-dimensional convolutional autoencoder is introduced to process selected indicators to remove redundancy and retain useful feature information. To fully capture the important degradation information within different stages in the feature sequences, a novel multi-scale attention feature fusion module is proposed to extract degradation features at different time scales. Considering the impact of degradation modes on RUL prediction, a dual-task prediction module based on no degradation mode labels is designed to obtain accurate RUL. Findings Comparative experiments and ablation studies on the PHM2012 bearing dataset verified the effectiveness of the proposed method. Furthermore, the rationality of the selected parameters is confirmed through model parameter analysis. Originality/value The novelty of the proposed method is that it not only provides prior knowledge but also further removes redundant information from prior knowledge. In addition, the distribution differences between the original features and their multi-scale convolution results are measured through Kullback–Leibler divergence as the attention scores, which allows the proposed method to focus on important information at different time scales.

Point spread function deconvolution using a convolutional autoencoder

A major issue in optical astronomical image analysis is the combined effect of the instrument’s point spread function (PSF) and the atmospheric seeing that blurs images and changes their shape in a way that is band and time-of-observation dependent. In this work we present a very simple neural network based approach to nonblind image deconvolution that relies on feeding a convolutional autoencoder (CAE) input images that have been preprocessed by convolution with the corresponding PSF and its regularized inverse, a method which is both conceptually simple and computationally less intensive. We also present here, a new approach for dealing with limited input dynamic range of neural networks compared to the dynamic range present in astronomical images. Published by the American Physical Society 2024

Time series forecasting of multiphase microstructure evolution using deep learning

Microstructure evolution, which plays a critical role in determining materials properties, is commonly simulated by the high-fidelity but computationally expensive phase-field method. To address this, we approximate microstructure evolution as a time series forecasting problem within the domain of deep learning. Our approach involves implementing a cost-effective surrogate model that accurately predicts the spatiotemporal evolution of microstructures, taking an example of spinodal decomposition in binary and ternary mixtures. Our surrogate model combines a convolutional autoencoder to reduce the dimensional representation of these microstructures with convolutional recurrent neural networks to forecast their temporal evolution. We use different variants of recurrent neural networks to compare their efficacy in developing surrogate models for phase-field predictions. On average, our deep learning framework demonstrates excellent accuracy and speedup relative to the “ground truth” phase-field simulations. We use quantitative measures to demonstrate how surrogate model predictions can effectively replace the phase-field timesteps without compromising accuracy in predicting the long-term evolution trajectory. Additionally, by emulating a transfer learning approach, our framework performs satisfactorily in predicting new microstructures resulting from alloy composition and physics unknown to the model. Therefore, our approach offers a useful data-driven alternative and accelerator to the materials microstructure simulation workflow.

Privacy-Preserving Detection of Tampered Radio-Frequency Transmissions Utilizing Federated Learning in LoRa Networks.

LoRa networks, widely adopted for low-power, long-range communication in IoT applications, face critical security concerns as radio-frequency transmissions are increasingly vulnerable to tampering. This paper addresses the dual challenges of privacy-preserving detection of tampered transmissions and the identification of unknown attacks in LoRa-based IoT networks. Leveraging Federated Learning (FL), our approach enables the detection of tampered RF transmissions while safeguarding sensitive IoT data, as FL allows model training on distributed devices without sharing raw data. We evaluated the performance of multiple FL-enabled anomaly-detection algorithms, including Convolutional Autoencoder Federated Learning (CAE-FL), Isolation Forest Federated Learning (IF-FL), One-Class Support Vector Machine Federated Learning (OCSVM-FL), Local Outlier Factor Federated Learning (LOF-FL), and K-Means Federated Learning (K-Means-FL). Using metrics such as accuracy, precision, recall, and F1-score, CAE-FL emerged as the top performer, achieving 97.27% accuracy and a balanced precision, recall, and F1-score of 0.97, with IF-FL close behind at 96.84% accuracy. Competitive performance from OCSVM-FL and LOF-FL, along with the comparable results of K-Means-FL, highlighted the robustness of clustering-based detection methods in this context. Visual analyses using confusion matrices and ROC curves provided further insights into each model's effectiveness in detecting tampered signals. This research underscores the capability of federated learning to enhance privacy and security in anomaly detection for LoRa networks, even against unknown attacks, marking a significant advancement in securing IoT communications in sensitive applications.

Reduced-Order Model of Coal Seam Gas Extraction Pressure Distribution Based on Deep Neural Networks and Convolutional Autoencoders

There has been extensive research on the partial differential equations governing the theory of gas flow in coal mines. However, the traditional Proper Orthogonal Decomposition–Radial Basis Function (POD-RBF) reduced-order algorithm requires significant computational resources and is inefficient when calculating high-dimensional data for coal mine gas pressure fields. To achieve the rapid computation of gas extraction pressure fields, this paper proposes a model reduction method based on deep neural networks (DNNs) and convolutional autoencoders (CAEs). The CAE is used to compress and reconstruct full-order numerical solutions for coal mine gas extraction, while the DNN is employed to establish the nonlinear mapping between the physical parameters of gas extraction and the latent space parameters of the reduced-order model. The DNN-CAE model is applied to the reduced-order modeling of gas extraction flow–solid coupling mathematical models in coal mines. A full-order model pressure field numerical dataset for gas extraction was constructed, and optimal hyperparameters for the pressure field reconstruction model and latent space parameter prediction model were determined through hyperparameter testing. The performance of the DNN-CAE model order reduction algorithm was compared to the POD-RBF model order reduction algorithm. The results indicate that the DNN-CAE method has certain advantages over the traditional POD-RBF method in terms of pressure field reconstruction accuracy, overall structure retention, extremum capture, and computational efficiency.