Single Convolutional Neural Network Research Articles

ADNet: A Real-Time Floating Algae Segmentation Using Distillation Network

The segmentation of floating algae is a hot topic in the field of marine environmental research. Given the vastness of coastal areas and complex environments, algae detection models must have both higher performance and lower deployment costs. However, relying solely on a single Convolutional Neural Network (CNN) or transformer structure fails to achieve this objective. In this paper, a novel real-time floating algae segmentation method using a distillation network (ADNet) is proposed, based on the RGB images. ADNet can effectively transfer the performance of the transformer-based teacher network to the CNN-based student model while preserving its lightweight design. Faced with complex marine environments, we introduce a novel Channel Purification Module (CPM) to simultaneously strengthen algae features and purify interference responses. Importantly, the CPM achieves this operation without increasing any learnable parameters. Moreover, considering the huge scale differences among algae targets in surveillance RGB images, we propose a lightweight multi-scale feature fusion network (L-MsFFN) to improve the student’s modeling ability across various scales. Additionally, to mitigate interference from low-level noises on higher-level semantics, a novel position purification module (PPM) is proposed. The PPM can achieve more accurate weight attention calculation between different pyramid levels, thereby enhancing the effectiveness of fusion. Compared to CNNs and transformers, our ADNet strikes an optimal balance between performance and speed. Extensive experimental results demonstrate that our ADNet achieves higher application performance in the field of floating algae monitoring tasks.

AI-assisted automatic MRI-based tongue volume evaluation in motor neuron disease (MND).

Motor neuron disease (MND) causes damage to the upper and lower motor neurons including the motor cranial nerves, the latter resulting in bulbar involvement with atrophy of the tongue muscle. To measure tongue atrophy, an operator independent automatic segmentation of the tongue is crucial. The aim of this study was to apply convolutional neural network (CNN) to MRI data in order to determine the volume of the tongue. A single triplanar CNN of U-Net architecture trained on axial, coronal, and sagittal planes was used for the segmentation of the tongue in MRI scans of the head. The 3D volumes were processed slice-wise across the three orientations and the predictions were merged using different voting strategies. This approach was developed using MRI datasets from 20 patients with 'classical' spinal amyotrophic lateral sclerosis (ALS) and 20 healthy controls and, in a pilot study, applied to the tongue volume quantification to 19 controls and 19 ALS patients with the variant progressive bulbar palsy (PBP). Consensus models with softmax averaging and majority voting achieved highest segmentation accuracy and outperformed predictions on single orientations and consensus models with union and unanimous voting. At the group level, reduction in tongue volume was not observed in classical spinal ALS, but was significant in the PBP group, as compared to controls. Utilizing single U-Net trained on three orthogonal orientations with consequent merging of respective orientations in an optimized consensus model reduces the number of erroneous detections and improves the segmentation of the tongue. The CNN-based automatic segmentation allows for accurate quantification of the tongue volumes in all subjects. The application to the ALS variant PBP showed significant reduction of the tongue volume in these patients and opens the way for unbiased future longitudinal studies in diseases affecting tongue volume.

SCTNet: Single-Branch CNN with Transformer Semantic Information for Real-Time Segmentation

Recent real-time semantic segmentation methods usually adopt an additional semantic branch to pursue rich long-range context. However, the additional branch incurs undesirable computational overhead and slows inference speed. To eliminate this dilemma, we propose SCTNet, a single branch CNN with transformer semantic information for real-time segmentation. SCTNet enjoys the rich semantic representations of an inference-free semantic branch while retaining the high efficiency of lightweight single branch CNN. SCTNet utilizes a transformer as the training-only semantic branch considering its superb ability to extract long-range context. With the help of the proposed transformer-like CNN block CFBlock and the semantic information alignment module, SCTNet could capture the rich semantic information from the transformer branch in training. During the inference, only the single branch CNN needs to be deployed. We conduct extensive experiments on Cityscapes, ADE20K, and COCO-Stuff-10K, and the results show that our method achieves the new state-of-the-art performance. The code and model is available at https://github.com/xzz777/SCTNet.

Deep learning for automatic prediction of early activation of treatment naïve non-exudative MNVs in AMD.

Around 30% of non-exudative macular neovascularizations(NE-MNVs) exudate within 2 years from diagnosis in patients with age-related macular degeneration(AMD).The aim of the study is to develop a deep learning classifier based on optical coherence tomography(OCT) and OCT angiography(OCTA) to identify NE-MNVs at risk of exudation. AMD patients showing OCTA and fluorescein angiography (FA) documented NE-MNV with a 2-years minimum imaging follow-up were retrospectively selected. Patients showing OCT B-scan-documented MNV exudation within the first 2 years formed the EX-GROUP while the others formed QU-GROUP.ResNet-101, Inception-ResNet-v2 and DenseNet-201 were independently trained on OCTA and OCT B-scan images. Combinations of the 6 models were evaluated with major and soft voting techniques. Eighty-nine (89) eyes of 89 patients with a follow-up of 5.7 ± 1.5 years were recruited(35 EX GROUP and 54 QU GROUP). Inception-ResNet-v2 was the best performing among the 3 single convolutional neural networks(CNNs).The major voting model resulting from the association of the 3 different CNNs resulted in improvement of performance both for OCTA and OCT B-scan (both significantly higher than human graders' performance). Soft voting model resulting from the combination of OCTA and OCT B-scan based major voting models showed a testing accuracy of 94.4%. Peripheral arcades and large vessels on OCTA enface imaging were more prevalent in QU GROUP. Artificial intelligence shows high performances in identifications of NE-MNVs at risk for exudation within the first 2 years of follow up, allowing better customization of follow up timing and avoiding treatment delay. Better results are obtained with the combination of OCTA and OCT B-scan image analysis.

State of charge estimation for Li-ion battery during dynamic driving process based on dual-channel deep learning methods and conditional judgement

The accurate estimation of the state of charge (SOC) is crucial for the safe and reliable operation of Li-ion batteries. In this paper, novel dual-channel deep learning methods were proposed for SOC estimation during dynamic driving cycles. For dual-channel deep learning methods, each channel of the deep learning methods could be a single convolutional neural network (CNN) layer, single long short-term memory (LSTM) layer or sequentially combined structure of CNN and LSTM. After evaluation, it could be found that dual-channel deep learning methods could achieve higher accuracy than conventional single-channel method, however, the model size and floating-point operations (FLOPs) could be either reduced or increased. Among all the dual-channel deep learning methods, CNN + LSTM-CNN + LSTM method could achieve the highest accuracy and the shortest training time with sacrifices in computational complexity. CNN–CNN + LSTM method could achieve higher accuracy than the single-channel deep learning method with reduced computational complexity and model size. As the error of SOC estimation is highly correlated with voltage signals, conditional judgement was then integrated with dual-channel deep learning method to obtain higher accuracy. After integrated with conditional judgement, the averaged error of SOC estimation could be reduced by more than 45% without a substantial sacrifice in computational time.

Fault Diagnosis of Fuel Cells by a Hybrid Deep Learning Network Fusing Characteristic Impedance

Establishing an accurate fault diagnosis system for the proton exchange membrane (PEM) fuel cell is essential for ensuring stable performance and delaying degradation. In this paper, a novel fault diagnosis method fusing characteristic impedance for the PEM fuel cell based on a hybrid deep learning network by combing the residual network (ResNet) and long short-term memory (LSTM) is proposed. Specifically, the characteristic impedance that can reflect internal dynamics loss is fused as the feature input alongside other commonly vehicular measurement signals and decoded to form a feature matrix. The feature matrix is then transferred to realize 25 categories of fault detection, including different degrees of membrane drying, flooding, and air starvation. The results showed the accuracy of ResNet-LSTM can reach 99.632% with a good balance of computational burden and is higher than that of a single LSTM, ResNet, and convolutional neural network, as well as traditional machine learning methods because such the hybrid structure can make full use of feature learning capability of ResNet and the time series analysis capability of LSTM. Moreover, the proposed hybrid framework is further validated and compared under different proportions of training samples, noise levels, and input signals.

Hybrid deep learning models for time series forecasting of solar power

Forecasting solar power production accurately is critical for effectively planning and managing renewable energy systems. This paper introduces and investigates novel hybrid deep learning models for solar power forecasting using time series data. The research analyzes the efficacy of various models for capturing the complex patterns present in solar power data. In this study, all of the possible combinations of convolutional neural network (CNN), long short-term memory (LSTM), and transformer (TF) models are experimented. These hybrid models also compared with the single CNN, LSTM and TF models with respect to different kinds of optimizers. Three different evaluation metrics are also employed for performance analysis. Results show that the CNN–LSTM–TF hybrid model outperforms the other models, with a mean absolute error (MAE) of 0.551% when using the Nadam optimizer. However, the TF–LSTM model has relatively low performance, with an MAE of 16.17%, highlighting the difficulties in making reliable predictions of solar power. This result provides valuable insights for optimizing and planning renewable energy systems, highlighting the significance of selecting appropriate models and optimizers for accurate solar power forecasting. This is the first time such a comprehensive work presented that also involves transformer networks in hybrid models for solar power forecasting.

Convolutional neural network intelligent fault diagnosis method for rotating machinery based on discriminant correlation analysis multi-domain feature fusion strategy

Aiming at solving the problems of limited training data, single input information, and limited diagnostic accuracy under the influence of strong background noise in fault diagnosis of rotating machinery, this paper proposes a fault diagnosis method based on the combination of discriminant correlation analysis (DCA) and convolutional neural network (CNN). Firstly, the original vibration signal is divided into several segments in the time domain, and the training data is directly processed by one CNN branch to extract multi-scale time domain features. Simultaneously, the divided data is subjected to discrete wavelet transform (DWT), and processed by another branch of CNN to extract multi-scale time-frequency features. Then, the DCA feature fusion mechanism is adopted to fuse the two-domain features extracted in the parallel branches to improve the model’ detection ability. Finally, the fused features are input into the deep CNN for training and learning to extract new features and output the classification results. Through the experimental analysis of two different types of data, the results show that the proposed method can be used for fault diagnosis of rotating machinery effectively. Compared with the single CNN network, the proposed method combines the multi-domain multi-scale feature extraction module with the DCA feature fusion module to enrich the feature information extraction ability. At the same time, the network performance is improved to get higher fault classification accuracy higher.

Development of a Detection System for Types of Weeds in Maize (Zea mays L.) under Greenhouse Conditions Using the YOLOv5 v7.0 Model

Corn (Zea mays L.) is one of the most important cereals worldwide. To maintain crop productivity, it is important to eliminate weeds that compete for nutrients and other resources. The eradication of these causes environmental problems through the use of agrochemicals. The implementation of technology to mitigate this impact is also a challenge. In this work, an artificial vision system was implemented based on the YOLOv5s (You Only Look Once) model, which uses a single convolutional neural network (CNN) that allows differentiating corn from four types of weeds, for which a mobile support structure was built to capture images. The performance of the trained model had a value of mAP@05 (mean Average Precision) at a threshold of 0.5 of 83.6%. A prediction accuracy of 97% and a mAP@05 of 97.5% were obtained for the maize class. For the weed classes, Lolium perenne, Sonchus oleraceus, Solanum nigrum, and Poa annua obtained an accuracy of 86%, 90%, 78%, and 74%, and a mAP@05 of 81.5%, 90.2%, 76.6% and 72.0%, respectively. The results are encouraging for the construction of a precision weeding system.

Deep learning CNN-APSO-LSSVM hybrid fusion model for feature optimization and gas-bearing prediction

Conventional machine learning (CML) methods have been successfully applied for gas reservoir prediction. Their prediction accuracy largely depends on the quality of the sample data; therefore, feature optimization of the input samples is particularly important. Commonly used feature optimization methods increase the interpretability of gas reservoirs; however, their steps are cumbersome, and the selected features cannot sufficiently guide CML models to mine the intrinsic features of sample data efficiently. In contrast to CML methods, deep learning (DL) methods can directly extract the important features of targets from raw data. Therefore, this study proposes a feature optimization and gas-bearing prediction method based on a hybrid fusion model that combines a convolutional neural network (CNN) and an adaptive particle swarm optimization-least squares support vector machine (APSO-LSSVM). This model adopts an end-to-end algorithm structure to directly extract features from sensitive multicomponent seismic attributes, considerably simplifying the feature optimization. A CNN was used for feature optimization to highlight sensitive gas reservoir information. APSO-LSSVM was used to fully learn the relationship between the features extracted by the CNN to obtain the prediction results. The constructed hybrid fusion model improves gas-bearing prediction accuracy through two processes of feature optimization and intelligent prediction, giving full play to the advantages of DL and CML methods. The prediction results obtained are better than those of a single CNN model or APSO-LSSVM model. In the feature optimization process of multicomponent seismic attribute data, CNN has demonstrated better gas reservoir feature extraction capabilities than commonly used attribute optimization methods. In the prediction process, the APSO-LSSVM model can learn the gas reservoir characteristics better than the LSSVM model and has a higher prediction accuracy. The constructed CNN-APSO-LSSVM model had lower errors and a better fit on the test dataset than the other individual models. This method proves the effectiveness of DL technology for the feature extraction of gas reservoirs and provides a feasible way to combine DL and CML technologies to predict gas reservoirs.

Single Shot Detector CNN and Deep Dilated Masks for Vision-Based Hand Gesture Recognition From Video Sequences

Single Shot Detector CNN and Deep Dilated Masks for Vision-Based Hand Gesture Recognition From Video Sequences

基于高光谱成像和卷积神经网络的欧李果实缺陷检测

In order to perform highly effective identification of external defects and increase the additional value of Cerasus Humilis fruits, this study used hyperspectral imaging technology to collect information on intact and defective Cerasus Humilis fruits. Based on the full transition spectrum, partial least squares discriminant analysis (PLS-DA) and back propagation neural networks (BPNN) were used to establish a discriminative model. The competitive adaptive reweighted sampling (CARS) was used to extract feature wavelengths, principal component analysis was used for data compression of single band images, BPNN and convolutional neural networks (CNN) were used for defect Cerasus Humilis fruits recognition of principal component images. The results showed that the overall detection accuracy of PLS-DA and BPNN models based on wavelength spectral information were 83.81% and 85.71%, respectively. BPNN was used to establish the calibration model based on the selected characteristic wavelengths by CARS, the accuracy rate was 90.47%. The classified accuracy of CNN model based on principal component images was 93.33%, which was obviously better than that of BPNN model at 83.81%. The research shows that the CNN model was successfully applied to the detection of Cerasus Humilis fruits defects using hyperspectral imaging. This study provides a theoretical basis for the development of fruit grading and sorting equipment.

An intelligent fault detection and diagnosis monitoring system for reactor operational resilience: Fault diagnosis

Various advanced reactor designs proposed in recent years envision deployment scenarios which feature reactor operations with significantly reduced staffing or even completely autonomous frameworks to reduce the operations and management costs of the plants. Many SMR and microreactor designs feature extended fuel cycles which limit inspection intervals, reduced access to critical components, and load-following capabilities that expose the reactor unit to different transients. Safe and reliable semi- or fully autonomous operations under these challenging operational regimes must be enabled through an on-line monitoring (OLM) system that effectively detects and diagnoses malfunctions in the reactor plant. This work presents the development of the Fault Diagnosis Module of the previously proposed data-driven OLM system for reactor operations: the Fault Detection and Diagnosis Monitoring System (FDDMS). The Fault Diagnosis Module monitors various sensor signatures from multiple systems and components at a nuclear power plant and accurately diagnoses the type and location of a malfunction. When integrated into the complete FDDMS methodology, the Fault Diagnosis Module can provide power transient dependent fault characterization by using separate convolutional neural network (CNN) models for steady state, ramping up in power, and ramping down in power operations, making the FDDMS especially applicable to load-following operational strategies. Two separate diagnosis approaches were explored in this paper for the Fault Diagnosis Module architecture: Hierarchical and End-to-End. The Hierarchical approach is a two-stage architecture in which the first stage uses a single CNN to classify the plant subsystem where the malfunction initiated. Subsequently in the second stage, a separate CNN is used for each subsystem to describe the specific fault type. The End-to-End approach uses a single CNN to directly classify the fault type from the overall list. To efficiently utilize computational resources, the hyperband intelligent hyperparameter optimization method to generate optimal CNN architectures. The diagnosis approaches were compared in terms of precision rate, recall rate, F1-score, and total accuracy for the possible fault types. Both diagnosis approaches produced good and satisfactory performance by generating total diagnosis accuracies above 99% on 17 different malfunction scenarios for all three power transient datasets. Additionally, robustness against noisy sensor data was tested, with models maintaining 99–100% accuracy at various levels of noise, and an illustrative real-time application of the methodology is provided.

Enhancing Stock Price Prediction Method Based on CNN-LSTM hybrid model

Stock price prediction means using the previous data to predict future trend. With the development of deep learning technologies in artificial intelligence area, many works try to use single CNN or single LSTM to make prediction. However, both these two models have their own weakness and strengths. In this paper, we try to make a CNN-LSTM hybrid model to achieve a better performance on stock price prediction. It proposes that the CNN-LSTM hybrid model predicts stock prices using RMSE to measure higher accuracy than a single CNN and a single LSTM model. The experiment shows a great ability on generalization and provide reliable results, which can apply to other stocks to make prediction and give suggestions on investing strategy. The CNN-LSTM model can play the respective advantages of CNN and LSTM model to improve the accuracy of stock price prediction and has better performance in both short-term prediction and long-term trend.

Enhanced accuracy for classification of video capsule endoscopy images using multiple deep learning convolutional neural networks

Background and aimsVideo capsule endoscopy (VCE) is widely used in the detection of abnormalities in the small intestine. However, there remains the challenge to correctly identify a limited number of possible abnormal images from tens of thousands of total images, and this impediment has limited the expansion of the technology. More recently artificial intelligence (AI) technology has been used in classifying VCE images from patients, but a clinical-grade diagnostic accuracy (greater than 99%) has not been achieved. MethodsThis study proposes a system for the automatic classification of a number of categories of unbounded VCE images with high accuracy by means of a transfer learning approach using multiple convolutional neural networks (CNNs). With this new approach, it is not necessary to implement image segmentation, so that the feature extraction becomes automatic and the existing models can be fine-tuned to obtain specific classifiers. ResultsOver 16,000 VCE gastrointestinal (GI) images from normal individuals including those with normal clean mucosa, the pylorus, the ileocecal valve, reduced mucosal view due to luminal contents and lymphangiectasia (a normal variant), and patients with five pathological states including angioectasia, bleeding, erosion/s, ulcers and foreign bodies, were obtained from a publicly available data set. These were used in building, testing and validating AI models for evaluating the diagnostic accuracy of our combined 17-CNN deep learning approach. Compared to a single CNN approach used by other research groups, our AI method, using 17 CNNs, achieved an overall diagnostic accuracy of 99.79%, with an accuracy of 100% for identifying bleeding and foreign bodies. The high accuracy was further demonstrated in the confusion matrices, precision, recall, and F1 score. ConclusionsWe have developed accurate AI deep learning models for unbounded VCE image classification of various medical conditions in medical practice.

Advanced Hybrid Model for Multi Paddy diseases detection using Deep Learning

INTRODUCTION: Rapid developments in deep learning (DL) techniques have made it possible to find and recognize objects in pictures. To create a network that is significantly more successful than a single CNN, GAN, RNN, etc., we can mix various neural network models (CNN, GAN, RNN).this combination is known as hybrid model. Hybrid model of deep leaning is give more accurately result for detection and identification of paddy diseases.
 OBJECTIVES: I have studies outcome of hybrid model 1(DCNN+SVM) and Hybrid model 2 (DCNN + Transfer Learning) to increase accuracy of Rice plant disease detection and classification. The Researched model detects multiple rice plant diseases and it is giving same result in multiple data sets.
 METHODS: The Proposed System have used Deep Learning Image Processing algorithm and neural Network Like DCNN ,SVM and Transfer Learning .The brand new model is DST where D stands for DCNN, S stands for SVM and T stands for transfer learning.
 RESULTS: The Researched DST model achieved 95% Training accuracy and 85% validation Accuracy. The Researched model detect multiple rice plant diseases and it is giving same result in multiple data set.
 CONCLUSION: The proposed model combined 2 existing model and developed hybrid model that a detect various rice plant diseases with better accuracy from available existing model.

A New Method to Predict Damage to Composite Structures Using Convolutional Neural Networks.

To reduce the cost of developing composite aeronautical structures, manufacturers and university researchers are increasingly using "virtual testing" methods. Then, finite element methods (FEMs) are intensively used to calculate mechanical behavior and to predict the damage to fiber-reinforced polymer (FRP) composites under impact loading, which is a crucial design aspect for aeronautical composite structures. But these FEMs require a lot of knowledge and a significant number of IT resources to run. Therefore, artificial intelligence could be an interesting way of sizing composites in terms of impact damage tolerance. In this research, the authors propose a methodology and deep learning-based approach to predict impact damage to composites. The data are both collected from the literature and created using an impact simulation performed using an FEM. The data augmentation method is also proposed to increase the data number from 149 to 2725. Firstly, a CNN model is built and optimized, and secondly, an aggregation of two CNN architectures is proposed. The results show that the use of an aggregation of two CNNs provides better performance than a single CNN. Finally, the aggregated CNN model prediction demonstrates the potential for CNN models to accelerate composite design by showing a 0.15 mm precision for all the length measurements, an average delaminated surface error of 56 mm2, and an error rate of 7% for the prediction of the presence of delamination.

Estimation of 3D anatomically précised hand poses using single shot corrective CNN

A crucial component of human-computer interaction is 3D hand posture assessment. The most recent advancements in computer vision have made estimating 3D hand positions simpler by using deep sensors. The main challenge still stems from unrealistic 3D hand poses because the existing models only use the training dataset to learn the kinematic rules, which is ambiguous, and it is a difficult task to estimate realistic 3D hand poses from datasets because they are not free from anatomical errors. The suggested model in this study is trained using a closed-form expression that encodes the biomechanical rules, thus it does not entirely reliant on the pictures from the annotated dataset. This work also used a Single Shot Detection and Correction convolutional neural network (SSDC-CNN) to handle the issues in imposing anatomically correctness from the architecture level. The ResNetPlus is implemented to improve representation capability with enhanced the efficiency of error back-propagation of the network. The datasets of the Yoga Mudras, like HANDS2017, and MSRA have been used to train and test the future model. As observed from the ground truth the previous hand models have many anatomical errors but, the proposed hand model is anatomically error free hand model compared to previous hand models. By considering the ground truth hand pose, the recommended hand model has shown good accuracy when compared to the state-of-art hand models.

Updates in Diagnostic Imaging for Infectious Keratitis: A Review

Infectious keratitis (IK) is among the top five leading causes of blindness globally. Early diagnosis is needed to guide appropriate therapy to avoid complications such as vision impairment and blindness. Slit lamp microscopy and culture of corneal scrapes are key to diagnosing IK. Slit lamp photography was transformed when digital cameras and smartphones were invented. The digital camera or smartphone camera sensor’s resolution, the resolution of the slit lamp and the focal length of the smartphone camera system are key to a high-quality slit lamp image. Alternative diagnostic tools include imaging, such as optical coherence tomography (OCT) and in vivo confocal microscopy (IVCM). OCT’s advantage is its ability to accurately determine the depth and extent of the corneal ulceration, infiltrates and haze, therefore characterizing the severity and progression of the infection. However, OCT is not a preferred choice in the diagnostic tool package for infectious keratitis. Rather, IVCM is a great aid in the diagnosis of fungal and Acanthamoeba keratitis with overall sensitivities of 66–74% and 80–100% and specificity of 78–100% and 84–100%, respectively. Recently, deep learning (DL) models have been shown to be promising aids for the diagnosis of IK via image recognition. Most of the studies that have developed DL models to diagnose the different types of IK have utilised slit lamp photographs. Some studies have used extremely efficient single convolutional neural network algorithms to train their models, and others used ensemble approaches with variable results. Limitations of DL models include the need for large image datasets to train the models, the difficulty in finding special features of the different types of IK, the imbalance of training models, the lack of image protocols and misclassification bias, which need to be overcome to apply these models into real-world settings. Newer artificial intelligence technology that generates synthetic data, such as generative adversarial networks, may assist in overcoming some of these limitations of CNN models.

MBC-Net: long-range enhanced feature fusion for classifying remote sensing images

PurposeClassification of remote sensing images (RSI) is a challenging task in computer vision. Recently, researchers have proposed a variety of creative methods for automatic recognition of RSI, and feature fusion is a research hotspot for its great potential to boost performance. However, RSI has a unique imaging condition and cluttered scenes with complicated backgrounds. This larger difference from nature images has made the previous feature fusion methods present insignificant performance improvements.Design/methodology/approachThis work proposed a two-convolutional neural network (CNN) fusion method named main and branch CNN fusion network (MBC-Net) as an improved solution for classifying RSI. In detail, the MBC-Net employs an EfficientNet-B3 as its main CNN stream and an EfficientNet-B0 as a branch, named MC-B3 and BC-B0, respectively. In particular, MBC-Net includes a long-range derivation (LRD) module, which is specially designed to learn the dependence of different features. Meanwhile, MBC-Net also uses some unique ideas to tackle the problems coming from the two-CNN fusion and the inherent nature of RSI.FindingsExtensive experiments on three RSI sets prove that MBC-Net outperforms the other 38 state-of-the-art (STOA) methods published from 2020 to 2023, with a noticeable increase in overall accuracy (OA) values. MBC-Net not only presents a 0.7% increased OA value on the most confusing NWPU set but also has 62% fewer parameters compared to the leading approach that ranks first in the literature.Originality/valueMBC-Net is a more effective and efficient feature fusion approach compared to other STOA methods in the literature. Given the visualizations of grad class activation mapping (Grad-CAM), it reveals that MBC-Net can learn the long-range dependence of features that a single CNN cannot. Based on the tendency stochastic neighbor embedding (t-SNE) results, it demonstrates that the feature representation of MBC-Net is more effective than other methods. In addition, the ablation tests indicate that MBC-Net is effective and efficient for fusing features from two CNNs.