Search Results (71)

This study focuses on predicting β-carotene content using hyperspectral images captured in the near-infrared (NIR) region during the drying process. Several machine learning models are compared, including Partial Least Squares Regression (PLSR), Stacked Autoencoders (SAEs) combined with Random Forest (RF), and Convolutional Neural Networks (CNNs) in three configurations: 1D-CNN, 2D-CNN, and 3D-CNN. The models are evaluated using R², Mean Absolute Error (MAE), and Root Mean Squared Error (RMSE). The PLSR model showed excellent results with R² = 0.97 for both training and testing, indicating minimal overfitting. The SAE-RF model also performed well, with R² values of 0.82 and 0.83 for training and testing, respectively, showing strong consistency. The CNN models displayed varying results: 1D-CNN achieved moderate performance, while 2D-CNN and 3D-CNN exhibited signs of overfitting, especially on testing data. Overall, the findings suggest that although CNNs are capable of capturing complex patterns, the PLSR and SAE-RF models deliver more reliable and robust predictions for β-carotene content in hyperspectral imaging. Full article

This paper presents the flow field reconstruction and prediction of powder fuel transport systems based on representative feature extraction from scattering images using deep learning techniques. A laboratory-built powder fuel supply system was used to conduct scattering spectroscopy experiments on boron-based fuel under various flow rate conditions. Based on the acquired scattering images, a prediction and reconstruction method was developed using a deep network framework composed of a Stacked Autoencoder (SAE), a Backpropagation Neural Network (BP), and a Long Short-Term Memory (LSTM) model. The proposed framework enables accurate classification and prediction of the dynamic evolution of flow structures based on learned representations from scattering images. Experimental results show that the feature vectors extracted by the SAE form clearly separable clusters in the latent space, leading to high classification accuracy under varying flow conditions. In the prediction task, the feature vectors predicted by the LSTM exhibit strong agreement with ground truth, with average mean square error, mean absolute error, and r-square values of 0.0027, 0.0398, and 0.9897, respectively. Furthermore, the reconstructed images offer a visual representation of the changing flow field, validating the model’s effectiveness in structure-level recovery. These results suggest that the proposed method provides reliable support for future real-time prediction of powder fuel mass flow rates based on optical sensing and imaging techniques. Full article

Single-pixel imaging (SPI) is a significant branch of computational imaging. Owing to the high sensitivity, low cost, and wide spectrum, it acquires extensive applications across various domains. Nevertheless, multiple measurements and long reconstruction time constrain its application. The application of neural networks has significantly improved the quality of reconstruction, but there is still a huge space for improvement in performance. SAE and Unet have different advantages in the field of SPI. However, there is no method that combines the advantages of these two networks for SPI reconstruction. Therefore, we propose the EMNP-SPI method for SPI reconstruction using SAE and Unet networks. The SAE makes use of the measurement dimension information and uses the group inverse to obtain the decoding matrix to enhance its generalization. The Unet uses different size convolution kernels and attention mechanisms to enhance feature extraction capabilities. Simulations and experiments confirm that our proposed enhanced multi-network prior method can significantly improve the quality of image reconstruction at low measurement rates. Full article

This paper presents a hybrid fast charging strategy for static wireless power transfer (WPT) systems that synergistically combines pulsed current and multi-stage current (MCM) modes to enable rapid yet battery-health-conscious electric vehicle (EV) charging, thereby promoting sustainable transportation. The proposed approach employs a memetic algorithm (MA) to dynamically optimize the charging parameters, achieving an optimal balance between speed and battery longevity while maintaining 90.78% system efficiency at the SAE J2954-standard 85 kHz operating frequency. A neural-network-based state of charge (SOC) estimator provides accurate real-time monitoring, complemented by MA-tuned PI control for enhanced resonance stability and adaptive pulsed current–MCM profiles for the optimal energy transfer. Simulations and experimental validation demonstrate faster charging compared to that using the conventional constant current–constant voltage (CC-CV) methods while effectively preserving the battery’s state of health (SOH)—a critical advantage that reduces the environmental impact of frequent battery replacements and minimizes the carbon footprint associated with raw material extraction and battery manufacturing. By addressing both the technical challenges of high-power WPT systems and the ecological imperative of battery preservation, this research bridges the gap between fast charging requirements and sustainable EV adoption, offering a practical solution that aligns with global decarbonization goals through optimized resource utilization and an extended battery service life. Full article

To improve the robustness of intrusion detection systems constructed using deep learning models, a method based on an auxiliary adversarial training WGAN (AuxAtWGAN) is proposed from the defender’s perspective. First, one-dimensional traffic data are downscaled and processed into two-dimensional image data via a stacked autoencoder (SAE), and mixed adversarial samples are generated using the fast gradient sign method (FGSM), Projected Gradient Descent (PGD) and Carlini and Wagner (C&W) adversarial attacks. Second, the improved WGAN with an integrated perceptual network module is trained with mixed training samples composed of mixed adversarial samples and normal samples. Finally, the adversary-trained AuxAtWGAN model is attached to the original model for adversary sample detection, and the detected adversary samples are removed and input into the original model to improve the robustness of the original model. The average attack success rate of the original convolutional neural network (CNN) model against multiple adversarial samples is 75.17%, and after using AuxAtWGAN, the average attack success rate of the adversarial attacks decreases to 27.56%; moreover, the detection accuracy of the original CNN model against normal samples is still 93.57%. The experiment proves that AuxAtWGAN improves the robustness of the original model. In addition, validation experiments are conducted by attaching the AuxAtWGAN model to the Long Short-Term Memory Network (LSTM) and Residual Network34 (ResNet) models, which prove that the proposed method has high generalization performance. Full article

The Rare Earth Extraction Process (REEP) model is difficult to accurately establish via the extraction mechanism method due to its high complexity. This paper proposes a multi-branch deep feature fusion network with SAE (SAE-MBDFFN) for modeling REEP. We first design a neural network with a multi-branch output structure to simulate the cascade REEP by introducing a multiscale feature fusion mechanism, which can simultaneously concatenate hidden features, original features, and inter-branch coupling features. In order to deal with insufficient labeled data during model training, we then adopt a stacked Sparse Auto-Encoder (SAE) technology to extract the hidden information of mass unlabeled data based on unsupervised learning. This technology can determine the initial parameters of SAE-MBDFFN by unsupervised pretraining. The design methodology of the network is well-founded. Experiments on industrial data indicate that the proposed method has the lowest initial loss value and a faster convergence rate in the fine-tuning stage than other comparison methods, while the prediction accuracy is better well. These results show the effectiveness of the proposed method. Full article

To address the limitations of traditional footstep vibration signal localization algorithms, such as limited accuracy, single feature extraction, and cumbersome parameter adjustment, a motion target localization method for step vibration signals based on deep learning is proposed. Velocity vectors are used to describe human motion and adapt it to the nonlinear motion and complex interactions of moving targets. In the feature extraction stage, a one-dimensional residual convolutional neural network is constructed to extract the time–frequency domain features of the signals, and a channel attention mechanism is introduced to enhance the model’s focus on different vibration sensor signal features. Furthermore, a bidirectional long short-term memory network is built to learn the temporal relationships between the extracted signal features of the convolution operation. Finally, regression operations are performed through fully connected layers to estimate the position and velocity vectors of the moving target. The dataset consists of footstep vibration signal data from six experimental subjects walking on four different paths and the actual motion trajectories of the moving targets obtained using a visual tracking system. Experimental results show that compared to WT-TDOA and SAE-BPNN, the positioning accuracy of our method has been improved by 37.9% and 24.8%, respectively, with a system average positioning error reduced to 0.376 m. Full article

In this paper, we perform neural cryptanalysis on five block ciphers: Data Encryption Standard (DES), Simplified DES (SDES), Advanced Encryption Standard (AES), Simplified AES (SAES), and SPECK. The block ciphers are investigated on three different deep learning-based attacks, Encryption Emulation (EE), Plaintext Recovery (PR), Key Recovery (KR), and Ciphertext Classification (CC) attacks. The attacks attempt to break the block ciphers in various cases, such as different types of plaintexts (i.e., block-sized bit arrays and texts), different numbers of round functions and quantity of training data, different text encryption methods (i.e., Word-based Text Encryption (WTE) and Sentence-based Text Encryption (STE)), and different deep learning model architectures. As a result, the block ciphers can be vulnerable to EE and PR attacks using a large amount of training data, and STE can improve the strength of the block ciphers, unlike WTE, which shows almost the same classification accuracy as the plaintexts, especially in a CC attack. Moreover, especially in the KR attack, the Recurrent Neural Network (RNN)-based deep learning model shows higher average Bit Accuracy Probability than the fully connected-based deep learning model. Furthermore, the RNN-based deep learning model is more suitable than the transformer-based deep learning model in the CC attack. Besides, when the keys are the same as the plaintexts, the KR attack can perfectly break the block ciphers, even if the plaintexts are randomly generated. Additionally, we identify that DES and SPECK32/64 applying two round functions are more vulnerable than those applying the single round function by performing the KR attack with randomly generated keys and randomly generated single plaintext. Full article

Multi-temporal Interferometric Synthetic Aperture Radar technique (MTInSAR) has emerged as a valuable tool for measuring ground motion in a wide area. However, interpreting displacement time series and identifying dangerous signals from millions of InSAR coherent targets is challenging. In this study, we propose a method combining stacked autoencoder (SAE) and convolutional neural network (CNN) to classify InSAR time series and ease the interpretation of movements. The InSAR time series are classified into five categories, including stable, linear, accelerating, deceleration, and phase unwrapping error (PUE). The accuracy of labeled samples reaches 95.1%, reflecting the performance of the proposed method. This method was applied to the InSAR results for Kunming extracted from 171 ascending Sentinel-1 images from January 2017 to September 2022. The classification map of the InSAR time series shows that stable coherent points dominate around 79.28% of the area, with linear patterns at 10.70%, decelerating at 5.30%, accelerating at 4.72%, and PUE patterns at 3.60%. The results demonstrate that this method can distinguish different ground motion features and detect nonlinear deformation signals on a large scale without human intervention. Full article

Grape quality and ripeness play a crucial role in producing exceptional wines with high-value characteristics, which requires an effective assessment of grape ripeness. The primary purpose of this research is to explore the possible application of visible–near-infrared spectral (Vis-NIR) technology for classifying the maturity stages of wine grapes based on quality indicators. The reflection spectra of Cabernet Sauvignon grapes were recorded using a spectrometer in the spectral range of 400 nm to 1029 nm. After measuring the soluble solids content (SSC), total acids (TA), total phenols (TP), and tannins (TN), the grape samples were categorized into five maturity stages using a spectral clustering method. A traditional supervised classification method, a support vector machine (SVM), and two deep learning techniques, namely stacked autoencoders (SAE) and one-dimensional convolutional neural networks (1D-CNN), were employed to construct a discriminant model and investigate the association linking grape maturity stages and the spectral responses. The spectral data went through three commonly used preprocessing methods, and feature wavelengths were extracted using a competitive adaptive reweighting algorithm (CARS). The spectral data model preprocessed via multiplicative scattering correction (MSC) outperformed the other two preprocessing methods. After preprocessing, a comparison was made between the discriminant models established with full and effective spectral data. It was observed that the SAE model, utilizing the feature spectrum, demonstrated superior overall performance. The classification accuracies of the calibration and prediction sets were 100% and 94%, respectively. This study showcased the dependability of combining Vis-NIR spectroscopy with deep learning methods for rapidly and accurately distinguishing the ripeness stage of grapes. It has significant implications for future applications in wine production and the development of optoelectronic instruments tailored to the specific needs of the winemaking industry. Full article

Remote sensing imagery involves capturing and examining details about the Earth’s surface from a distance, often using satellites, drones, or other aerial platforms. It offers useful data with which to monitor and understand different phenomena on Earth. Vehicle detection and classification play a crucial role in various applications, including traffic monitoring, urban planning, and environmental analysis. Deep learning, specifically convolutional neural networks (CNNs), has revolutionized vehicle detection in remote sensing. This study designs an improved Chimp optimization algorithm with a DL-based vehicle detection and classification (ICOA-DLVDC) technique on RSI. The presented ICOA-DLVDC technique involves two phases: object detection and classification. For vehicle detection, the ICOA-DLVDC technique applies the EfficientDet model. Next, the detected objects can be classified by using the sparse autoencoder (SAE) model. To optimize the SAE’s hyperparameters effectively, we introduce an ICOA which streamlines the parameter tuning process, accelerating convergence and enhancing the overall performance of the SAE classifier. An extensive set of experiments has been conducted to highlight the improved vehicle classification outcomes of the ICOA-DLVDC technique. The simulation values demonstrated the remarkable performance of the ICOA-DLVDC approach compared to other recent techniques, with a maximum accuracy of 99.70% and 99.50% on the VEDAI dataset and ISPRS Postdam dataset, respectively. Full article

Turbofan engines are known as the heart of the aircraft. The turbofan’s health state determines the aircraft’s operational status. Therefore, the equipment monitoring and maintenance of the engine is an important part of ensuring the healthy and stable operation of the aircraft, and it is vital to monitor the remaining useful life (RUL) of the engine. The monitored data of turbofan engines have high dimensions and a long time span, which cause difficulties in predicting the remaining useful life of the engine. This paper proposes a residual life prediction model based on Autoencoder and a Temporal Convolutional Network (TCN). Among them, Autoencoder is used to reduce the dimension of the data and extract features from the engine monitoring data. The TCN network is trained on the obtained low-dimensional data to predict the remaining useful life. The model mentioned in this article is verified on the NASA public data set (C-MAPSS) and compared with common machine learning methods and other deep neural networks. The SAE-TCN model achieved better scores on the FD001 independent testing data set with an RMSE of 18.01 and a score of 161. The average relative error of the model relative to other common learning models is 0.9499 in RMSE and 0.2656 in Scoring Function. The experimental results show that the model proposed in this paper performs the best in the evaluation, and this conclusion has important implications for engine health. Full article

Load disaggregation determines appliance-level energy consumption unintrusively from aggregated consumption measured by a single meter. Deep neural networks have been proven to have great potential in load disaggregation. In this article, a temporal convolution network, mainly consisting of residual blocks with bidirectional dilated convolution, the GeLu activation function, and multihead attention, is proposed to improve the prediction accuracy of individual appliances. Bidirectional dilated convolution is applied to enlarge the receptive field and effectively extract load features from historical and future information. Meanwhile, GeLU is introduced into the residual structure to overcome the “dead state” issue of traditional ReLU. Furthermore, multihead attention aims to improve the prediction accuracy by giving different weights according to the importance of different-level load features. The proposed model is validated using the REDD and UK-DALE datasets. Among six existing neural networks, the experimental results demonstrate that the proposed algorithm achieves the least average errors when disaggregating four appliances in terms of mean absolute error (MAE) and signal aggregate error (SAE), respectively, reduced by $22.33\%$ and $60.58\%$ compared with the model with the second-best performance on the REDD dataset. Additionally, the proposed algorithm shows superior results in identifying the on/off state in four appliances from the UK-DALE dataset. Full article

Accurately assessing the intensity of pain from facial expressions captured in videos is crucial for effective pain management and critical for a wide range of healthcare applications. However, in uncontrolled environments, detecting facial expressions from full left and right profiles remains a significant challenge, and even the most advanced models for recognizing pain levels based on facial expressions can suffer from declining performance. In this study, we present a novel model designed to overcome the challenges posed by full left and right profiles—Sparse Autoencoders for Facial Expressions-based Pain Assessment (SAFEPA). Our model utilizes Sparse Autoencoders (SAE) to reconstruct the upper part of the face from the input image, and feeds both the original image and the reconstructed upper face into two pre-trained concurrent and coupled Convolutional Neural Networks (CNNs). This approach gives more weight to the upper part of the face, resulting in superior recognition performance. Moreover, SAFEPA’s design leverages CNNs’ strengths while also accommodating variations in head poses, thus eliminating the need for face detection and upper-face extraction preprocessing steps needed in other models. SAFEPA achieves high accuracy in recognizing four levels of pain on the widely used UNBC-McMaster shoulder pain expression archive dataset. SAFEPA is extended for facial expression recognition, where we show it to outperform state-of-the-art models in recognizing seven facial expressions viewed from five different angles, including the challenging full left and right profiles, on the Karolinska Directed Emotional Faces (KDEF) dataset. Furthermore, the SAFEPA system is capable of processing BioVid Heat Pain datasets with an average processing time of 17.82 s per video (5 s in length), while maintaining a competitive accuracy compared to other state-of-the-art pain detection systems. This experiment demonstrates its applicability in real-life scenarios for monitoring systems. With SAFEPA, we have opened new possibilities for accurate pain assessment, even in challenging situations with varying head poses. Full article

In order to characterize the flow behaviors of SAE 5137H steel, isothermal compression tests at the temperatures of 1123 K, 1213 K, 1303 K, 1393 K, and 1483 K, and the strain rates of 0.01 s⁻¹, 0.1 s⁻¹, 1 s⁻¹, and 10 s⁻¹ were performed using a Gleeble 3500 thermo-mechanical simulator. The analysis results of true stress-strain curves show that the flow stress decreases with temperature increasing and strain rate decreasing. In order to accurately and efficiently characterize the complex flow behaviors, the intelligent learning method backpropagation–artificial neural network (BP-ANN) was combined with the particle swarm optimization (PSO), namely, the PSO-BP integrated model. Detailed comparisons of the semi-physical model with improved Arrhenius-Type, BP-ANN, and PSO-BP integrated model for the flow behaviors of SAE 5137H steel in terms of generative ability, predictive ability, and modeling efficiency were presented. The comparison results show that the PSO-BP integrated model has the best comprehensive ability, BP-ANN is the second, and semi-physical model with improved Arrhenius-Type is the lowest. It indicates that the PSO-BP integrated model can accurately describe the flow behaviors of SAE 5137H steel. Full article