Experimental Comparison Research Articles

Using Anchor-Free Object Detectors to Detect Surface Defects

Due to the numerous disadvantages that come with having anchors in the detection process, a lot of researchers have been concentrating on the design of object detectors that do not rely on anchors. In this work, we use anchor-free object detectors in the field of computer vision for surface defect detection. First, we constructed a surface defect detection dataset about real wind turbine blades, which was supplemented with several methods due to the lack of natural data. Next, we used a number of popular anchor-free detectors (CenterNet, FCOS, YOLOX-S, and YOLOV8-S) to detect surface defects in this blade dataset. After experimental comparison, YOLOV8-S demonstrated the best detection performance, with a high accuracy (79.55%) and a short detection speed (9.52 fps). All the upcoming experiments are predicated on it. Third, we examined how the attention mechanism added to various YOLOV8-S model positions affected the two datasets—our blade dataset and the NEU dataset—and discovered that the insertion methods on the two datasets are the same when focusing on comprehensive performance. Lastly, we carried out a significant amount of experimental comparisons.

Experimental Investigation of the Impact of Mixed Wettability on Pore-Scale Fluid Displacement: A Microfluidic Study.

Understanding rock wettability is crucial across various fields including hydrology, subsurface fluid storage and extraction, and environmental sciences. In natural subsurface formations like carbonate and shale, mixed wettability is frequently observed, characterized by heterogeneous regions at the pore scale that exhibit both hydrophilic (water-wet) and hydrophobic (oil-wet) characteristics. Despite its common occurrence, the impact of mixed wettability on immiscible fluid displacement at the pore scale remains poorly understood, creating a gap in effective modeling and prediction of fluid behavior in porous media. The primary objective of this study was to investigate how mixed wettability affects pore-scale fluid displacement dynamics, utilizing microfluidic devices designed to replicate rock-like structures with varied wettability properties. Current techniques for achieving mixed wettability within microfluidic devices often struggle with spatial control and resolution, limiting their accuracy. To address this limitation, a novel approach was employed that combined photolithography and molecular vapor deposition of perfluorodecyltrichlorosilane to precisely and selectively modify wettability within specific pore regions, achieving a mixed wettability distribution correlated with pore size for the first time. The experimental setup included five identical micromodels representing distinct wetting conditions, which were initially saturated with air and subsequently flooded by water. By systematically varying the ratio of hydrophilic to hydrophobic areas, we covered a range from fully hydrophilic to fully hydrophobic and intermediate mixed wettability configurations. Comparative displacement experiments revealed that pore-level mixed wettability has a pronounced effect on fluid displacement behavior, influencing the injection time, spatial invasion patterns, and dynamic pressure profiles. Results indicated that both the injection time and dynamic pressure decreased with an increase in the hydrophilic area fraction. Each wettability configuration displayed unique sequences of pore-filling events, emphasizing the critical role of the wettability distribution in influencing displacement dynamics. While mixed wettability exhibited a clear monotonic effect on invasion time and dynamic pressure, saturation behavior was notably nonmonotonic. Interestingly, mixed wettability scenarios with relatively medium to high hydrophilic fractions demonstrated enhanced overall sweep efficiency compared to the hydrophobic case and reduced the bypassed gas phase relative to the hydrophilic case. However, inefficiently distributed mixed wet zones were found to reduce the sweep efficiency. These findings highlight the critical influence of mixed wettability in fluid displacement processes, with significant implications for applications in oil recovery, CO2 sequestration, and other subsurface energy technologies.

Mapping Lithology with Hybrid Attention Mechanism–Long Short-Term Memory: A Hybrid Neural Network Approach Using Remote Sensing and Geophysical Data

Remote sensing provides an efficient roadmap in geological analysis and interpretation. However, some challenges arise when remote sensing techniques are integrated with machine learning in geological surveys. Factors including irregular spatial distribution, sample imbalance, interclass resemblances, regolith, and geochemical similarities impede geological feature diagnosis, interpretation, and identification across varied remote sensing datasets. To address these limitations, a hybrid-attention-integrated long short-term memory (LSTM) network is employed to diagnose, interpret, and identify lithological feature representations in a remote sensing-based geological analysis using multisource data fusion. The experimental design integrates varied datasets including Sentinel-2A, Landsat-9, ASTER, ALOS PALSAR DEM, and Bouguer anomaly gravity data. The proposed model incorporates a hybrid attention mechanism (HAM) comprising channel and spatial attention submodules. HAM utilizes an adaptive technique that merges global-average-pooled features with max-pooled features, enhancing the model’s accuracy in identifying lithological units. Additionally, a channel separation operation is employed to allot refined channel features into clusters based on channel attention maps along the channel dimension. The comprehensive analysis of results from comparative extensive experiments demonstrates HAM-LSTM’s state-of-the-art performance, outperforming existing attention modules and attention-based models (ViT, SE-LSTM, and CBAM-LSTM). Comparing HAM-LSTM to baseline LSTM, the HAM module’s integrated configurations equip the proposed model to better diagnose and identify lithological units, thereby increasing the accuracy by 3.69%.

Cross-dataset micro-expression identification based on facial ROIs contribution quantification

Abstract In this work, we propose a novel cross-dataset micro-expression identification method based on facial regions of interest contribution quantification, where the training samples are from the source dataset and the test samples are from the target dataset. Specifically, a micro-expression video clip (i.e., a sample) is first sampled at intervals to capture an image sequence consisting of multiple consecutive video frames. Then, locate and crop the facial area in the image frame, and perform facial landmark detection on the cropped area. Next calculate the optical-flow field of the image sequence, and extract the improved Main Directional Mean Optical-flow feature. Finally, the learned group sparse model is used as the classifier to predict the label information of the unmarked target samples, so as to complete the category recognition of micro-expressions. Extensive cross-dataset comparative experiments on four spontaneous micro-expression datasets, CASME, CASME II, SMIC-HS, and SAMM, show that the recognition strategy proposed in this work is effective. Compared with several other advanced recognition methods, our method has better cross-dataset recognition performance, not only the classification accuracy is higher, but also the classification stability is better when faced with different target datasets and different micro-expression categories. It can be applied in many fields such as clinical diagnosis, interrogation, negotiation, and national security.

Class-Imbalanced Machinery Fault Diagnosis Using Heterogeneous Data Fusion Support Tensor Machine

The monitoring signals of bearings from single-source sensor often contain limited information for characterizing various working condition, which may lead to instability and uncertainty of the class-imbalanced intelligent fault diagnosis. On the other hand, the vectorization of multi-source sensor signals may not only generate high-dimensional vectors, leading to increasing computational complexity and overfitting problems, but also lose the structural information and the coupling information. This paper proposes a new method for class-imbalanced fault diagnosis of bearing using support tensor machine (STM) driven by heterogeneous data fusion. The collected sound and vibration signals of bearings are successively decomposed into multiple frequency band components to extract various time-domain and frequency-domain statistical parameters. A third-order heterogeneous feature tensor is designed based on multi sensors, frequency band components and statistical parameters. STM based intelligent model is constructed to preserve the structural information of the third-order heterogeneous feature tensor for bearing fault diagnosis. A series of comparative experiments verify the advantages of the proposed method. Conflict of Interest Statement The authors declare no conflicts of interest.

Real-time economic following velocity and gear planning for commercial vehicles based on the bi-level optimization

ABSTRACT Current Predictive Adaptive Cruise Control (PACC) systems pose significant challenges. These include poor collaborative optimisation between economic velocity and the powertrain system, conservative energy-saving strategies and underutilisation of map data due to real-time constraints. A computationally efficient collaborative optimisation method is proposed for economic following velocity and gear shifting based on Bi-level optimisation theory. Based on the Bi-level theory, economic following velocity and gear shifting are decoupled and hierarchically planned. The lower-level subproblem constructs the objective function based on a dynamic-weight safety distance model within the Model Predictive Control (MPC) framework, solving the economic following velocity by the Continuous Generalized Minimal Residual Method (C/GMRES). The upper-level subproblem solves the economic gear shifting by the optimal control law based on the economic following velocity. Finally, comparative experiments are conducted between the proposed PACC algorithm, a standard Adaptive Cruise Control (ACC) algorithm, and a benchmark Dynamic Programming-Model Predictive Control (DP-MPC) PACC algorithm. The results show that the proposed PACC algorithm reduces fuel consumption by approximately 5.76% compared to the ACC. Compared to the benchmark algorithm, the proposed PACC algorithm also demonstrates significantly improved computational efficiency while achieving comparable energy savings.

Multi-source domain generalization tool wear prediction based on wide convolution weighted antagonism

Abstract While traditional deep learning models achieve high accuracy in predicting tool wear under consistent working conditions, actual production processes frequently involve varying conditions due to different processing methods. The wear data of different working conditions have a large difference in distribution, so that the wear signal of milling cutter trained in one working condition can only predict the wear value of the working condition, which will cause a large waste of material and manpower for actual production. Therefore, a domain generalization method utilizing wide deep convolutional neural networks and weighted antagonistic multi-source domain generalization is introduced for tool wear prediction, and multi-source domain wide depth convolutional weighted adversarial domain generalization model (WDCNNWAL) model is constructed. After filtering and de-noising the original data, it is input into WDCNN for automatic feature learning, and the resulting multi-source feature signals are input into the prediction model and discriminator respectively to obtain corresponding loss values. Furthermore, Wasserstein distance is used to measure the probability distribution distance of multi-source feature signals as a weighted value combined with the discriminator loss function, and the final loss value is obtained by adding together. Finally, the tool wear prediction model is set up using a backpropagation neural network. Various milling wear experimental data along with the NASA tool wear public dataset were utilized to assess the predictive performance of the trained model. The model achieves a high level of generalization in tool wear prediction, with average root mean squared error (RMSE) of 0.0599 and 0.2075, as well as average R 2 of 0.9037 and 0.9196 on the NASA and self-collecting datasets respectively. To validate its exceptional precision in generalization performance, comparative experiments were conducted with other data-driven methods resulting in over a 50% reduction in average RMSE and more than a 40% increase in average R 2.

Study on the Influence of Scaling‐Up in the Transesterification Process for the Synthesis of Biodegradable Monoglycerides of Fatty Acids

ABSTRACTFatty acids and monoglycerides from vegetable oils or animal fats are crucial in various industries. Monoglycerides, with their hydrophilic and hydrophobic properties, act as effective non‐ionic surfactants for oil‐in‐water emulsions. Scaling‐up the synthesis of biodegradable non‐ionic surfactants from natural sources presents challenges in maintaining physicochemical properties. This study investigates the scale‐up synthesis of biodegradable non‐ionic surfactants using sunflower oil and glycerol, focusing on properties like kinematic viscosity, acid number, and active matter content. Comparative experiments were conducted using different reactor sizes and configurations, evaluating the effects of reactor size, length‐to‐diameter (L/D) ratio, and mixer type. The results indicate that larger reactors generally reduce kinematic viscosity and active matter content. The geometric characteristics of the reactor significantly impact surfactant properties with variations in the RD/RH ratio affecting the active matter content and changes in the MPD/RD ratio influencing the water number. Based on our findings, the highest active matter content was achieved with an RD/RH ratio of 0.49 and an MPD/RD ratio of 0.95. A strong correlation was found between the active matter content and water solubility. The study underscores the importance of reactor design in optimizing surfactant properties when scaling‐up.

Axial-UNet++ Power Line Detection Network Based on Gated Axial Attention Mechanism

The segmentation and recognition of power lines are crucial for the UAV-based inspection of overhead power lines. To address the issues of class imbalance, low sample quantity, and long-range dependency in images, a specialized semantic segmentation network for power line segmentation called Axial-UNet++ is proposed. Firstly, to tackle the issue of long-range dependencies in images and low sample quantity, a gated axial attention mechanism is introduced to expand the receptive field and improve the capture of relative positional biases in small datasets, thereby proposing a novel feature extraction module termed axial-channel local normalization module. Secondly, to address the imbalance in training samples, a new loss function is developed by combining traditional binary cross-entropy loss with focal loss, enhancing the precision of image semantic segmentation. Lastly, ablation and comparative experiments on the PLDU and Mendeley datasets demonstrate that the proposed model achieves 54.7% IoU and 80.1% recall on the PLDU dataset, and 79.3% IoU and 93.1% recall on the Mendeley dataset, outperforming other listed models. Additionally, robustness experiments show the adaptability of the Axial-UNet++ model under extreme conditions and the augmented image dataset used in this study has been open sourced.

A Temporal Convolutional Neural Network Fusion Attention Mechanism Runoff Prediction Model Based on Dynamic Decomposition Reconstruction Integration Processing

Accurate and reliable runoff forecasting is of great significance for hydropower station operation and watershed water resource allocation. However, various complex factors, such as climate conditions and human activities, constantly affect the formation of runoff. Runoff data under changing environments exhibit highly nonlinear, time-varying, and stochastic characteristics, which undoubtedly pose great challenges to runoff prediction. Under this background, this study ingeniously merges reconstruction integration technology and dynamic decomposition technology to propose a Temporal Convolutional Network Fusion Attention Mechanism Runoff Prediction method based on dynamic decomposition reconstruction integration processing. This method uses the Temporal Convolutional Network to extract the cross-temporal nonlinear characteristics of longer runoff data, and introduces attention mechanisms to capture the importance distribution and duration relationship of historical temporal features in runoff prediction. It integrates a decomposition reconstruction process based on dynamic classification and filtering, fully utilizing decomposition techniques, reconstruction techniques, complexity analysis, dynamic decomposition techniques, and neural networks optimized by automatic hyperparameter optimization algorithms, effectively improving the model’s interpretability and precision of prediction accuracy. This study used historical monthly runoff datasets from the Pingshan Hydrological Station and Yichang Hydrological Station for validation, and selected eight models including the LSTM model, CEEMDAN-TCN-Attention model, and CEEMDAN-VMD-LSTM-Attention (DDRI) for comparative prediction experiments. The MAE, RMSE, MAPE, and NSE indicators of the proposed model showed the best performances, with test set values of 1007.93, 985.87, 16.47, and 0.922 for the Pingshan Hydrological Station and 1086.81, 1211.18, 17.20, and 0.919 for the Yichang Hydrological Station, respectively. The experimental results indicate that the fusion model generated through training has strong learning ability for runoff temporal features and the proposed model has obvious advantages in overall predictive performance, stability, correlation, comprehensive accuracy, and statistical testing.

A multi-view learning approach with diffusion model to synthesize FDG PET from MRI T1WI for diagnosis of Alzheimer's disease.

This study presents a novel multi-view learning approach for machine learning (ML)-based Alzheimer's disease (AD) diagnosis. A diffusion model is proposed to synthesize the fluorodeoxyglucose positron emission tomography (FDG PET) view from the magnetic resonance imaging T1 weighted imaging (MRI T1WI) view and incorporate two synthesis strategies: one-way synthesis and two-way synthesis. To assess the utility of the synthesized views, we use multilayer perceptron (MLP)-based classifiers with various combinations of the views. The two-way synthesis achieves state-of-the-art performance with a structural similarity index measure (SSIM) at 0.9380 and a peak-signal-to-noise ratio (PSNR) at 26.47. The one-way synthesis achieves an SSIM at 0.9282 and a PSNR at 23.83. Both synthesized FDG PET views have shown their effectiveness in improving diagnostic accuracy. This work supports the notion that ML-based cross-domain data synthesis can be a useful approach to improve AD diagnosis by providing additional synthesized disease-related views for multi-view learning. We propose a diffusion model with two strategies to synthesize fluorodeoxyglucose positron emission tomography (FDG PET) from magnetic resonance imaging T1 weighted imaging (MRI T1WI). We raise multi-view learning with MRl T1Wl and synthesized FDG PET for Alzheimer's disease (AD) diagnosis. We provide a comprehensive experimental comparison for the synthesized FDG PET view. The feasibility of synthesized FDG PET view in AD diagnosis is validated with various experiments. We demonstrate the ability of synthesized FDG PET to enhance the performance of machine learning-based AD diagnosis.

LULC-SegNet: Enhancing Land Use and Land Cover Semantic Segmentation with Denoising Diffusion Feature Fusion

Deep convolutional networks often encounter information bottlenecks when extracting land object features, resulting in critical geometric information loss, which impedes semantic segmentation capabilities in complex geospatial backgrounds. We developed LULC-SegNet, a semantic segmentation network for land use and land cover (LULC), which integrates features from the denoising diffusion probabilistic model (DDPM). This network enhances the clarity of the edge segmentation, detail resolution, and the visualization and accuracy of the contours by delving into the spatial details of the remote sensing images. The LULC-SegNet incorporates DDPM decoder features into the LULC segmentation task, utilizing machine learning clustering algorithms and spatial attention to extract continuous DDPM semantic features. The network addresses the potential loss of spatial details during feature extraction in convolutional neural network (CNN), and the integration of the DDPM features with the CNN feature extraction network improves the accuracy of the segmentation boundaries of the geographical features. Ablation and comparison experiments conducted on the Circum-Tarim Basin Region LULC Dataset demonstrate that the LULC-SegNet improved the LULC semantic segmentation. The LULC-SegNet excels in multiple key performance indicators compared to existing advanced semantic segmentation methods. Specifically, the network achieved remarkable scores of 80.25% in the mean intersection over union (MIOU) and 93.92% in the F1 score, surpassing current technologies. The LULC-SegNet demonstrated an IOU score of 73.67%, particularly in segmenting the small-sample river class. Our method adapts to the complex geophysical characteristics of remote sensing datasets, enhancing the performance of automatic semantic segmentation tasks for land use and land cover changes and making critical advancements.

A motor imagery classification model based on hybrid brain-computer interface and multitask learning of electroencephalographic and electromyographic deep features

ObjectiveExtracting deep features from participants’ bioelectric signals and constructing models are key research directions in motor imagery (MI) classification tasks. In this study, we constructed a multimodal multitask hybrid brain-computer interface net (2M-hBCINet) based on deep features of electroencephalogram (EEG) and electromyography (EMG) to effectively accomplish motor imagery classification tasks.MethodsThe model first used a variational autoencoder (VAE) network for unsupervised learning of EEG and EMG signals to extract their deep features, and subsequently applied the channel attention mechanism (CAM) to select these deep features and highlight the advantageous features and minimize the disadvantageous ones. Moreover, in this study, multitask learning (MTL) was applied to train the 2M-hBCINet model, incorporating the primary task that is the MI classification task, and auxiliary tasks including EEG reconstruction task, EMG reconstruction task, and a feature metric learning task, each with distinct loss functions to enhance the performance of each task. Finally, we designed module ablation experiments, multitask learning comparison experiments, multi-frequency band comparison experiments, and muscle fatigue experiments. Using leave-one-out cross-validation(LOOCV), the accuracy and effectiveness of each module of the 2M-hBCINet model were validated using the self-made MI-EEMG dataset and the public datasets WAY-EEG-GAL and ESEMIT.ResultsThe results indicated that compared to comparative models, the 2M-hBCINet model demonstrated good performance and achieved the best results across different frequency bands and under muscle fatigue conditions.ConclusionThe 2M-hBCINet model constructed based on EMG and EEG data innovatively in this study demonstrated excellent performance and strong generalization in the MI classification task. As an excellent end-to-end model, 2M-hBCINet can be generalized to be used in EEG-related fields such as anomaly detection and emotion analysis.

Rapid Vehicle Trajectory Prediction Based on Multi-Attention Mechanism for Fusing Multimodal Information

Trajectory prediction plays a crucial role in autonomous driving tasks, as accurately and rapidly predicting the future trajectories of traffic participants can significantly enhance the safety and robustness of autonomous driving systems. This paper presents a novel trajectory prediction model that follows the encoder–decoder paradigm, achieving precise and rapid predictions of future vehicle trajectories by efficiently aggregating the spatiotemporal and interaction information of agents in traffic scenarios. We propose an agent–agent interaction information extraction module based on a sparse graph attention mechanism, which enables the efficient aggregation of interaction information between agents. Additionally, we introduce a non-autoregressive query generation method that accelerates the model inference speed by generating the decoding queries in parallel. Comparative experiments with the existing advanced algorithms show that our method improves the multimodal trajectory prediction metrics for the Minimum Average Displacement Error (minADE), the Minimum Final Displacement Error (minFDE), and the Miss Rate (MR) by an average of 9.1%, 11.8%, and 14.6%, respectively, while the inference time is only 33.7% of the average time taken by the other algorithms. Finally, we demonstrate the effectiveness of the various modules proposed in this paper through ablation studies.

A Comprehensive Proteomic and Bioinformatic Analysis of Human Spinal Cord Injury Plasma Identifies Proteins Associated with the Complement Cascade and Liver Function as Potential Prognostic Indicators of Neurological Outcome.

Spinal cord injury (SCI) is a major cause of disability, with complications postinjury often leading to lifelong health issues with the need for extensive treatment. Neurological outcome post-SCI can be variable and difficult to predict, particularly in incompletely injured patients. The identification of specific SCI biomarkers in blood may be able to improve prognostics in the field. This study has utilized proteomic and bioinformatic methodologies to investigate differentially expressed proteins in plasma samples across human SCI cohorts with the aim of identifying candidate prognostic biomarkers and biological pathway alterations that relate to neurological outcome. Blood samples were taken, following informed consent, from American Spinal Injury Association impairment scale (AIS) grade C "improvers" (those who experienced an AIS grade improvement) and "nonimprovers" (no AIS change) and AIS grade A and D at <2 weeks ("acute") and ∼3 months ("subacute") postinjury. The total protein concentration from each sample was extracted, with pooled samples being labeled and nonpooled samples treated with ProteoMiner™ beads. Samples were then analyzed using two 4-plex isobaric tag for relative and absolute quantification (iTRAQ) analyses and a label-free experiment for comparison before quantifying with mass spectrometry. Data are available via ProteomeXchange with identifiers PXD035025 and PXD035072 for the iTRAQ and label-free experiments, respectively. Proteomic datasets were analyzed using OpenMS (version 2.6.0). R (version 4.1.4) and, in particular, the R packages MSstats (version 4.0.1) and pathview (version 1.32.0) were used for downstream analysis. Proteins of interest identified from this analysis were further validated by enzyme-linked immunosorbent assay. The data demonstrated proteomic differences between the cohorts, with the results from the iTRAQ approach supporting those of the label-free analysis. A total of 79 and 87 differentially abundant proteins across AIS and longitudinal groups were identified from the iTRAQ and label-free analyses, respectively. Alpha-2-macroglobulin, retinol-binding protein 4 (RBP4), serum amyloid A1, peroxiredoxin 2 (PRX-2), apolipoprotein A1, and several immunoglobulins were identified as biologically relevant and differentially abundant, with potential as individual candidate prognostic biomarkers of neurological outcome. Bioinformatics analyses revealed that the majority of differentially abundant proteins were components of the complement cascade and most interacted directly with the liver. Many of the proteins of interest identified using proteomics were detected only in a single group and therefore have potential as binary (present or absent) biomarkers, RBP4 and PRX-2 in particular. Additional investigations into the chronology of these proteins and their levels in other tissues (cerebrospinal fluid in particular) are needed to better understand the underlying pathophysiology, including any potentially modifiable targets. Pathway analysis highlighted the complement cascade as being significant across groups of differential functional recovery.

Impact of Storage on In Vitro Permeation and Mucoadhesion Setup Experiments Using Swine Nasal Mucosa.

Intranasal topical administration offers a promising route for local and systemic drug delivery, with in vitro permeation and mucoadhesion studies often using porcine models. However, the impact of storage on mucosal integrity after the procedure remains unaddressed. This study aimed to standardize the preparation process and evaluated whether storage of porcine nasal mucosa impairs its integrity and permeability for experimental comparisons. Additionally, an optimized in vitro mucoadhesion experiment using texture analyzer equipment was investigated. Porcine nasal mucosa was subjected to different storage conditions ("fresh"; refrigerated at 4°C for 24h and 48h, and frozen at -20°C for two or three weeks) and assessed using optical and transmission electron microscopy. In vitro permeation assays were performed in a Franz-type vertical diffusion system with lidocaine hydrochloride (LDC). In vitro mucoadhesion assays were conducted using fresh nasal mucosa and a commercial nasal topical formulation using TA.XT. Plus texture analyzer. The variables involved (probe speed, contact time, and application force) in assessing mucoadhesive capacity (maximum mucoadhesive force Fmax and work of mucoadhesion Wmuc) were optimized using a Central Composite Design. Fresh tissues showed no alterations in histological arrangement or in the ultrastructure of adherence junctions. Stored tissues exhibited histological disorganization, reduced thickness, and loss of epithelial integrity. LDC permeability increased in storage tissues (p < 0.05). Contact force had a positive effect on Fmax and Wmuc (p < 0.0001), with a minimum required value of 0.48 N. Variations in contact time and probe speed did not affect the responses (p > 0.05). In conclusion, the preparation technique was adequate to maintain mucosa integrity for permeability studies. However, storing the mucosa at 4 or -20°C overestimated LDC permeation, which could mislead critical data for formulation development. Therefore, the use of fresh mucosa is recommended to ensure more reliable results. For in vitro mucoadhesion assays, a minimum contact force of 0.48N is required for optimal responses.

Marine propulsion shafting-oriented status detection algorithm for uncertainty time-varying

The marine propulsion shafting is the lifeline of navigation. The marine propulsion shafting vibration signals tend to contain rich connotations that suggest the shafting status. In this paper, aiming at the difficulty in the capture of time dependence in uncertainty time-varying signals of marine propulsion shafting, the concept of Degree of Membership (DM) in fuzzy mathematics is introduced into Recurrent Neural Network (RNN), called Degree of Membership Recurrent Neural Network (DM-RNN). A marine propulsion shafting status detection algorithm based on DM-RNN is proposed. Specifically, the mathematical expressions for forward propagation and back propagation of DM-RNN are theoretically derived. The self-propulsion test scheme based on a full-scale boat is designed. Actual vibration signals under multiple working conditions are collected. The progressiveness of the present algorithm under shafting vibration signals of the full-scale boat is verified by the experiment. The results of the two comparison experiments indicate that the present algorithm significantly improved the detection accuracy compared with the classical ANNs, up to 93.25% ± 0.71%. Furthermore, a DM function with a simple form and fewer parameters was more conducive to the high detection accuracy of DM-RNN.

LSYOLO: an algorithm for linear scan PCB defect detection

Abstract Aiming at the current problems that different defects in linear scan PCB have scale differences and some defects have high similarity with the background, which are difficult to localize and classify, an expanded receptive field PCB defect detection algorithm is proposed to be applied to the defect detection of linear scan circuit boards. The expanded receptive field module (ERFM) is used in the backbone of YOLOv8 to replace C2f, which can avoid information loss and gridding artifacts while obtaining better contextual information to improve the detection performance of defects with high background similarity. Then, the spatial selective feature pyramid (SSFPN) is used as the FPN to enhance the model’s ability to detect defects at different scales while reducing the model performance requirements by utilizing the information of the spatial dimensions of the feature maps to fuse the feature maps at different scales. Wise-intersection over union is used as the bounding box loss function, and slide loss is used as the classification loss function to enhance the model’s focus on difficult-to-localize and difficult-to-classify samples. Comparison experiments are conducted on a linear scan printed circuit board dataset, and the experimental results show that the improved model obtains a significant improvement in the detection performance of defects with high detection difficulty; the average precision of the overall defects is improved by 9.6%, the number of model parameters is reduced by 40%, the amount of computation is reduced by 20%, and the size of the model weights file is only 3.64MB. Detecting defects in linear scan PCB is more efficient and lighter than other algorithms.

Monte Carlo simulation of recrystallization and grain growth in Annealed Al–4% Cu alloy: Validation through experimental microstructure comparison

The study of recrystallization and grain growth in annealed metals has been extensively investigated using various computational approaches, with Monte Carlo (MC) simulations proving particularly effective. This research highlights the application of the MC method in simulating microstructures that closely resemble experimentally observed ones. The experimental microstructures were obtained from a 50% cold-worked Al–4% Cu (Duralumin) alloy, annealed at different conditions. The Metropolis algorithm was employed to simulate the microstructures using a two-dimensional Potts model on a square lattice. The simulated results strongly correlate with the actual microstructures, validating the effectiveness of the MC method in exploring grain growth phenomena. This study underscores the potential of MC simulations as a robust tool for investigating recrystallization and grain evolution in annealed metals.

Divergent Adsorption Regulation in Metal-Organic Frameworks for Highly Efficient CF4/C2F6 Separation.

The efficient removal of low-concentration components from homologous mixtures is often hampered by the co-directional effect of traditional thermodynamic regulation approaches, typically leading to a trade-off between adsorption capacity and selectivity. Focusing this challenge on the critical task of purifying perfluorocarbons in electronics industry, a divergent regulation strategy is reported that significantly improves the separation efficiency of low-concentration hexafluoroethane (C2F6) from tetrafluoromethane (CF4). This approach involves the selective shielding of open metal sites and the modulation of channel geometry within an electron-deficient ligand-based pore environment, thereby facilitating a C2F6 dense-packing accommodation mode while weakening the CF4 affinity due to the reduced host-guest interactions. Simultaneously enhanced C2F6 adsorption and reduced CF4 adsorption are achieved, resulting in record-high low-pressure C2F6 uptake and C2F6/CF4 selectivity. Comprehensive insights into the unique separation mechanism are illustrated through a combination of solid-state MAS nuclear magnetic resonance(SSNMR), molecular simulations, and meticulously designed comparative experiments. As a result, benchmark C2F6/CF4 separation performance is achieved, as demonstrated by the unprecedented electronic-grade (over 99.999%) CF4 productivity (401 L kg-1) obtained from an industrially relevant C2F6/CF4 (3:97) mixture, as well as the excellent water/air/heat stability and recyclability.