Multi-scale Convolutional Neural Network Research Articles

A flow rate estimation method for gas–liquid two-phase flow based on filter-enhanced convolutional neural network

Accurate estimation of flow rate in gas–liquid two-phase flow is crucial for various industrial processes. How to accurately estimate flow rate remains a challenging problem. Previously, deep learning-based methods focused on a few human-set points with single task learning. In addition, the data were not denoised. In this study, a flow rate estimation method based on a filter-enhanced convolutional neural network (FECNN) is proposed for gas–liquid two-phase flow. The method leverages multimodal data from a Venturi tube and an electrical capacitance tomography (ECT) sensor as input, utilizing multilayer perceptron (MLP) to fuse data. Subsequently, a learnable filter module is employed to attenuate noise adaptively, followed by multiscale convolutional neural network (MSCNN) extraction of flow rate features at different scales. Finally, the method enables estimate each single-phase flow rate simultaneously through multi-task learning (MTL). The adaptive noise attenuation capabilities of the learnable filter module are demonstrated, and the ability of the proposed MSCNN to capture multiscale flow rate features through multiple comparative experiments is shown. Additionally, a qualitative comparison with recent flow rate estimation methods is provided. Overall, this study demonstrates the effectiveness and superiority of the proposed FECNN in flow rate estimation.

A Multi-Scale Convolutional Neural Network with Self-Knowledge Distillation for Bearing Fault Diagnosis

Efficient bearing fault diagnosis not only extends the operational lifespan of rolling bearings but also reduces unnecessary maintenance and resource waste. However, current deep learning-based methods face significant challenges, particularly due to the scarcity of fault data, which impedes the models’ ability to effectively learn parameters. Additionally, many existing methods rely on single-scale features, hindering the capture of global contextual information and diminishing diagnostic accuracy. To address these challenges, this paper proposes a Multi-Scale Convolutional Neural Network with Self-Knowledge Distillation (MSCNN-SKD) for bearing fault diagnosis. The MSCNN-SKD employs a five-stage architecture. Stage 1 uses wide-kernel convolution for initial feature extraction, while Stages 2 through 5 integrate a parallel multi-scale convolutional structure to capture both global contextual information and long-range dependencies. In the final two stages, a self-distillation process enhances learning by allowing deep-layer features to guide shallow-layer learning, improving performance, especially in data-limited scenarios. Extensive experiments on multiple datasets validate the model’s high diagnostic accuracy, computational efficiency, and robustness, demonstrating its suitability for real-time industrial applications in resource-limited environments.

Novel Approach in Vegetation Detection Using Multi-Scale Convolutional Neural Network

Vegetation segmentation plays a crucial role in accurately monitoring and analyzing vegetation cover, growth patterns, and changes over time, which in turn contributes to environmental studies, land management, and assessing the impact of climate change. This study explores the potential of a multi-scale convolutional neural network (MSCNN) design for object classification, specifically focusing on vegetation detection. The MSCNN is designed to integrate multi-scale feature extraction and attention mechanisms, enabling the model to capture both fine and coarse vegetation patterns effectively. Moreover, the MSCNN architecture integrates multiple convolutional layers with varying kernel sizes (3 × 3, 5 × 5, and 7 × 7), enabling the model to extract features at different scales, which is vital for identifying diverse vegetation patterns across various landscapes. Vegetation detection is demonstrated using three diverse datasets: the CamVid dataset, the FloodNet dataset, and the multispectral RIT-18 dataset. These datasets present a range of challenges, including variations in illumination, the presence of shadows, occlusion, scale differences, and cluttered backgrounds, which are common in real-world scenarios. The MSCNN architecture allows for the integration of information from multiple scales, facilitating the detection of diverse vegetation types under varying conditions. The performance of the proposed MSCNN method is rigorously evaluated and compared against state-of-the-art techniques in the field. Comprehensive experiments showcase the effectiveness of the approach, highlighting its robustness in accurately segmenting and classifying vegetation even in complex environments. The results indicate that the MSCNN design significantly outperforms traditional methods, achieving a remarkable global accuracy and boundary F1 score (BF score) of up to 98%. This superior performance underscores the MSCNN’s capability to enhance vegetation detection in imagery, making it a promising tool for applications in environmental monitoring and land use management.

MCNN-CMCA: A multiscale convolutional neural networks with cross-modal channel attention for physiological signal-based mental state recognition

Human mental state recognition (MSR) has significant implications for human-machine interactions. Although mental state recognition models based on single-modality signals, such as electroencephalogram (EEG) or peripheral physiological signals (PPS), have achieved encouraging progress, methods leveraging multimodal physiological signals still need to be explored. In this study, we present MCNN-CMCA, a generic model that employs multiscale convolutional neural networks (CNNs) with cross-modal channel attention to realize physiological signals-based MSR. Specifically, we first design an innovative cross-modal channel attention mechanism that adaptively adjusting the weights of each signal channel, effectively learning both intra-modality and inter-modality correlation and expanding the channel information to the depth dimension. Additionally, the study utilizes multiscale temporal CNNs for obtaining short-term and long-term time-frequency features across different modalities. Finally, the multimodal fusion module integrates the representations of all physiological signals and the classification layer implements sparse connections by setting the mask weights to 0. We evaluate the proposed method on the SEED-VIG, DEAP, and self-made datasets, achieving superior results compared to existing state-of-the-art methods. Furthermore, we conduct ablation studies to demonstrate the effectiveness of each component in the MCNN-CMCA and show the use of multimodal physiological signals outperforms single-modality signals.

A Multi-Scale CNN for Transfer Learning in sEMG-Based Hand Gesture Recognition for Prosthetic Devices

A Multi-Scale CNN for Transfer Learning in sEMG-Based Hand Gesture Recognition for Prosthetic Devices

Fault prediction of rolling bearings using a multi-scale convolutional neural network with parallel BiLSTM for noise environment

Fault prediction of rolling bearings using a multi-scale convolutional neural network with parallel BiLSTM for noise environment

Wheat growth stage identification method based on multimodal data

Accurate identification of crop growth stages is a crucial basis for implementing effective cultivation management. With the development of deep learning techniques in image understanding, research on intelligent real-time recognition of crop growth stages based on RGB images has garnered significant attention. However, the small differences and high similarity in crop morphological characteristics during the transition between adjacent growth stages pose challenges for accurate identification. To address this issue, this study proposes a multi-scale convolutional neural network model, termed MultiScalNet-Wheat (MSN-W), which enhances the algorithm's ability to learn complex features by utilizing multi-scale convolution and attention mechanisms. This model extracts key information from redundant data to identify winter wheat growth stages in complex field environments. Experimental results show that the MSN-W model achieves a recognition accuracy of 97.6 %, outperforming typical convolutional neural network models such as VGG19, ResNet50, MobileNetV3, and DenseNet. To further address the difficulty in recognizing growth stages during transition periods, where canopy morphological features are highly similar and show small differences, this paper introduces an innovative approach by incorporating sequential environmental data related to wheat growth stages. By extracting these features and performing multi-modal collaborative inference, a multi-modal feature-based wheat growth stage recognition model, termed MultiModalNet-Wheat (MMN-W), is constructed on the basis of the MSN-W model. Experimental results indicate that the MMN-W model achieves a recognition accuracy of 98.5 %, improving by 0.9 % over the MSN-W model. Both the MSN-W and MMN-W models provide accurate methods for observing wheat growth stages, thereby supporting the scientific management of winter wheat at different growth stages.

A novel approach for tool-narayanaswamy-moynihan model parameter extraction using multi-scale neural model

The accurate determination of parameters in the Tool-Narayanaswamy-Moynihan (TNM) model, which describes the viscoelastic behavior of glass-forming materials, is crucial for predicting material responses through various thermal histories.Traditional methods rely heavily on curve-fitting techniques; however, these often fail due to noise in the data. Furthermore, traditional methods are computationally intensive and prone to inaccuracies, particularly when dealing with complex datasets or when the initial parameter guesses are far from optimal; also, they require a skilled personnel.In this study, we propose the application of a multi-scale convolutional neural network (MCNN) as a machine learning approach to address these challenges. The MCNN model is trained on a comprehensive simulated dataset encompassing a wide range of TNM parameters, allowing it to learn intricate patterns and dependencies within the data that are difficult to capture with conventional methods. Our results show that the MCNN significantly improves the accuracy of the parameter estimations for β and x across the entire spectrum of tested conditions, achieving performance that is not only comparable to, but often surpasses, traditional curve-fitting methods. Furthermore, the MCNN demonstrates superior robustness when initial parameter estimates are suboptimal or when the dataset exhibits significant noise. Although the prediction accuracy for the activation energy Δh∗ and the pre-exponential factor log(A) was somewhat lower, the method still provides valuable estimates that can be refined with supplementary techniques.This work highlights the potential of machine learning approaches like MCNN to revolutionize the parameter extraction process in complex physical models, reducing the reliance on manual curve-fitting and providing a more automated, scalable solution. We also analyze the primary sources of prediction errors in the MCNN outputs and offer insights into future improvements, including model architecture refinements and the integration of additional physical constraints. Our findings suggest that this approach can be extended to other domains where similar models are employed, paving the way for broader applications of machine learning in materials science.

DeepDualEnhancer: A Dual-Feature Input DNABert Based Deep Learning Method for Enhancer Recognition.

Enhancers are cis-regulatory DNA sequences that are widely distributed throughout the genome. They can precisely regulate the expression of target genes. Since the features of enhancer segments are difficult to detect, we propose DeepDualEnhancer, a DNABert-based method using a multi-scale convolutional neural network, BiLSTM, for enhancer identification. We first designed the DeepDualEnhancer method based only on the DNA sequence input. It mainly consists of a multi-scale Convolutional Neural Network, and BiLSTM to extract features by DNABert and embedding, respectively. Meanwhile, we collected new datasets from the enhancer-promoter interaction field and designed the method DeepDualEnhancer-genomic for inputting DNA sequences and genomic signals, which consists of the transformer sequence attention. Extensive comparisons of our method with 20 other excellent methods through 5-fold cross validation, ablation experiments, and an independent test demonstrated that DeepDualEnhancer achieves the best performance. It is also found that the inclusion of genomic signals helps the enhancer recognition task to be performed better.

Multi-stream encoder and multi-layer comparative learning network for fluid classification based on logging data via wavelet threshold denoising

In recent years, the importance of fluid classification in oil and gas exploration has become increasingly evident. However, the inherent complexity of logging data and noise pose significant challenges to this task. To this end, this paper proposes a wavelet threshold denoising-based multi-stream encoder combined with multi-level comparison learning (LogMEC-MCL) framework for fluid classification. The framework begins with comprehensive noise reduction, utilizing wavelet threshold denoising to preprocess the data. It then extracts global temporal features by incorporating attention gated recurrent units within the multi-stream encoder. In parallel, multi-scale convolutional neural networks capture local spatial information, ensuring a more complete understanding of the data. To further improve the discriminative power of the extracted features, the framework includes two contrastive learning modules: instance-level contrastive learning and temporal contrastive learning. These components work together to refine feature differentiation, particularly in challenging cases. Additionally, the framework introduces a custom-designed loss function that combines cross-entropy loss with contrastive loss, thereby optimizing the classification performance. The proposed model was rigorously evaluated using a real-world logging dataset from the Tarim Basin in China. The experimental results demonstrate that LogMEC-MCL consistently outperforms current state-of-the-art models on two test datasets, achieving maximum classification accuracies of 95.70% and 95.50%, respectively.

A Wear Condition Warning Method for Wind Turbine Gearbox Based on Oil Online Monitoring Using Learnable Multiscale Convolutional Neural Network

A Wear Condition Warning Method for Wind Turbine Gearbox Based on Oil Online Monitoring Using Learnable Multiscale Convolutional Neural Network

Multi-Scale convolutional neural network for finger vein recognition

With the continuous advancement of science and technology, an increasing number of deep learning methods are being applied in the field of finger vein recognition to describe the structural characteristics of finger veins. However, some deep learning methods fail to adequately extract longer texture features. during the feature extraction process, resulting in a decrease in the uniqueness of extracted finger vein features. Additionally, these methods tend to extract global information while neglecting the importance of local texture information. To address the aforementioned issues, this paper introduces a multiscale convolution network (MCNet) model based on finger vein structure. On one hand, a multiscale feature extraction (MFE) model based on the rectangular and square convolution kernels are employed to extract structural information from finger veins and to simultaneously enhance the features of longer texture features. On the other hand, the paper introduces a cross-information fusion attention (CFA) block that combines spatial and channel information, in order to enhance local details information and the network’s ability to extract vein patterns. The experimental results on the public datasets FV-USM, SDUMLA-HMT, and HKPU validate the effectiveness of MCNet with the recognition rates of 99.86%, 99.11%, and 99.15% respectively.

Multi-Scale Convolutional Neural Networks optimized by elite strategy dung beetle optimization algorithm for encrypted traffic classification

Multi-Scale Convolutional Neural Networks optimized by elite strategy dung beetle optimization algorithm for encrypted traffic classification

Optimization and imputation of acoustic absorption sequence prediction for transversely isotropic porous materials: A MultiScaleCNN-Transformer-PSO model approach

To address the limitations of traditional methods in real-time prediction and adaptive optimization of sound absorption coefficients for transversely isotropic porous materials. This study introduces a method combining a Multi-Scale Convolutional Neural Network (MultiScaleCNN) and a Transformer model to improve real-time prediction and optimization of sound absorption coefficients in transversely isotropic porous materials. Using a validated transfer matrix method based on the Biot model and acoustic analysis via COMSOL 6.2, a database of sound absorption curves was created. MultiScaleCNN extracts multi-scale features, which the Transformer model uses for sequence-to-sequence modeling with a self-attention mechanism, effectively managing complex dependencies. The model achieves high accuracy in full-mask prediction, closely aligning with actual values. Additionally, Particle Swarm Optimization (PSO) optimizes sound absorption coefficients in the 1000–2000 Hz range, meeting the fitness function’s criteria. For partial-mask imputation, the model uses adjacent frequency data and features to enhance accuracy. Results show that the MultiScaleCNN-Transformer-PSO model offers a robust, data-driven approach for predicting and optimizing sound absorption coefficients, advancing the field.

Center-Highlighted Multiscale CNN for Classification of Hyperspectral Images

Hyperspectral images (HSIs) capture a wide range of spectral features across multiple bands of light, from visible to near-infrared. Hyperspectral image classification technology enables researchers to accurately identify and analyze the composition and distribution of surface materials. Current mainstream deep learning methods typically use block sampling to capture spatial features for the model. However, this approach can affect classification results due to the influence of neighboring features within the sample block. To improve the model’s focus on the center of the sampling block, this study proposes a center highlight with multiscale CNN for hyperspectral image classification (CHMSC). The network utilizes an automatic channel selector (Auto-CHS) to fully consider every channel feature and capture the correlation between the channels. Then, CHMSC enhances the model’s ability to concentrate on the central features of the sampling block utilizing structures such as the center highlight. Finally, before outputting the prediction results, an SENet is employed to further refine the features and learn associate interactions between different scales of spatial features and spectral features. Experimental results from three hyperspectral datasets validate the effectiveness of the proposed method. Specifically, when 15 samples from each class are selected for training, CHMSC achieves the highest overall accuracy (OA) of 90.05%, 92.78%, and 90.15% on the three datasets, outperforming other methods with increases of more than 3.11%, 1.8%, and 2.01% in OA, respectively.

Precise grading of non-muscle invasive bladder cancer with multi-scale pyramidal CNN

The grading of non-muscle invasive bladder cancer (NMIBC) continues to face challenges due to subjective interpretations, which affect the assessment of its severity. To address this challenge, we are developing an innovative artificial intelligence (AI) system aimed at objectively grading NMIBC. This system uses a novel convolutional neural network (CNN) architecture called the multi-scale pyramidal pretrained CNN to analyze both local and global pathology markers extracted from digital pathology images. The proposed CNN structure takes as input three levels of patches, ranging from small patches (e.g., 128×128\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$128 \ imes 128$$\\end{document}) to the largest size patches (512×512\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$512 \ imes 512$$\\end{document}). These levels are then fused by random forest (RF) to estimate the severity grade of NMIBC. The optimal patch sizes and other model hyperparameters are determined using a grid search algorithm. For each patch size, the proposed system has been trained on 32K patches (comprising 16K low-grade and 16K high-grade samples) and subsequently tested on 8K patches (consisting of 4K low-grade and 4K high-grade samples), all annotated by two pathologists. Incorporating light and efficient processing, defining new benchmarks in the application of AI to histopathology, the ShuffleNet-based AI system achieved notable metrics on the testing data, including 94.25% ± 0.70% accuracy, 94.47% ± 0.93% sensitivity, 94.03% ± 0.95% specificity, and a 94.29% ± 0.70% F1-score. These results highlight its superior performance over traditional models like ResNet-18. The proposed system’s robustness in accurately grading pathology demonstrates its potential as an advanced AI tool for diagnosing human diseases in the domain of digital pathology.

Fire spread prediction model based on multi-scale convolutional neural network

Fire spread prediction model based on multi-scale convolutional neural network

Optimized deep learning networks for accurate identification of cancer cells in bone marrow

Radiologists utilize pictures from X-rays, magnetic resonance imaging, or computed tomography scans to diagnose bone cancer. Manual methods are labor-intensive and may need specialized knowledge. As a result, creating an automated process for distinguishing between malignant and healthy bone is essential. Bones that have cancer have a different texture than bones in unaffected areas. Diagnosing hematological illnesses relies on correct labeling and categorizing nucleated cells in the bone marrow. However, timely diagnosis and treatment are hampered by pathologists' need to identify specimens, which can be sensitive and time-consuming manually. Humanity's ability to evaluate and identify these more complicated illnesses has significantly been bolstered by the development of artificial intelligence, particularly machine, and deep learning. Conversely, much research and development is needed to enhance cancer cell identification—and lower false alarm rates. We built a deep learning model for morphological analysis to solve this problem. This paper introduces a novel deep convolutional neural network architecture in which hybrid multi-objective and category-based optimization algorithms are used to optimize the hyperparameters adaptively. Using the processed cell pictures as input, the proposed model is then trained with an optimized attention-based multi-scale convolutional neural network to identify the kind of cancer cells in the bone marrow. Extensive experiments are run on publicly available datasets, with the results being measured and evaluated using a wide range of performance indicators. In contrast to deep learning models that have already been trained, the total accuracy of 99.7% was determined to be superior.

A New Scene Sensing Model Based on Multi-Source Data from Smartphones

Smartphones with integrated sensors play an important role in people’s lives, and in advanced multi-sensor fusion navigation systems, the use of individual sensor information is crucial. Because of the different environments, the weights of the sensors will be different, which will also affect the method and results of multi-source fusion positioning. Based on the multi-source data from smartphone sensors, this study explores five types of information—Global Navigation Satellite System (GNSS), Inertial Measurement Units (IMUs), cellular networks, optical sensors, and Wi-Fi sensors—characterizing the temporal, spatial, and mathematical statistical features of the data, and it constructs a multi-scale, multi-window, and context-connected scene sensing model to accurately detect the environmental scene in indoor, semi-indoor, outdoor, and semi-outdoor spaces, thus providing a good basis for multi-sensor positioning in a multi-sensor navigation system. Detecting environmental scenes provides an environmental positioning basis for multi-sensor fusion localization. This model is divided into four main parts: multi-sensor-based data mining, a multi-scale convolutional neural network (CNN), a bidirectional long short-term memory (BiLSTM) network combined with contextual information, and a meta-heuristic optimization algorithm.

The Research of Intra-Pulse Modulated Signal Recognition of Radar Emitter under Few-Shot Learning Condition Based on Multimodal Fusion

Radar radiation source recognition is critical for the reliable operation of radar communication systems. However, in increasingly complex electromagnetic environments, traditional identification methods face significant limitations. These methods often struggle with high noise levels and diverse modulation types, making it difficult to maintain accuracy, especially when the Signal-to-Noise Ratio (SNR) is low or the available training data are limited. These difficulties are further intensified by the necessity to generalize in environments characterized by a substantial quantity of noisy, low-quality signal samples while being constrained by a limited number of desirable high-quality training samples. To more effectively address these issues, this paper proposes a novel approach utilizing Model-Agnostic Meta-Learning (MAML) to enhance model adaptability in few-shot learning scenarios, allowing the model to quickly learn with limited data and optimize parameters effectively. Furthermore, a multimodal fusion neural network, DCFANet, is designed, incorporating residual blocks, squeeze and excitation blocks, and a multi-scale CNN, to fuse I/Q waveform data and time–frequency image data for more comprehensive feature extraction. Our model enables more robust signal recognition, even when the signal quality is severely degraded by noise or when only a few examples of a signal type are available. Testing on 13 intra-pulse modulated signals in an Additive White Gaussian Noise (AWGN) environment across SNRs ranging from −20 to 10 dB demonstrated the approach’s effectiveness. Particularly, under a 5−way5−shot setting, the model achieves high classification accuracy even at −10dB SNR. Our research underscores the model’s ability to address the key challenges of radar emitter signal recognition in low-SNR and data-scarce conditions, demonstrating its strong adaptability and effectiveness in complex, real-world electromagnetic environments.