Single Convolutional Neural Network Research Articles

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.

A Voice Disease Detection Method Based on MFCCs and Shallow CNN

The incidence rate of voice diseases is increasing year by year. The use of software for remote diagnosis is a technical development trend and has important practical value. Among voice diseases, common diseases that cause hoarseness include spasmodic dysphonia, vocal cord paralysis, vocal nodule, and vocal cord polyp. This paper presents a voice disease detection method that can be applied in a wide range of clinical. We cooperated with Xiangya Hospital of Central South University to collect voice samples from 352 different patients. The Mel Frequency Cepstrum Coefficient (MFCC) parameters are extracted as input features to describe the voice in the form of data. An innovative model combining MFCC parameters and single convolution layer CNN is proposed for fast calculation and classification. The highest accuracy we achieved was 92%, it is fully ahead of the original research results and internationally advanced. And we use advanced voice function assessment databases (AVFAD) to evaluate the generalization ability of the method we proposed, which achieved an accuracy rate of 98%. Experiments on clinical and standard datasets show that for the pathological detection of voice diseases, our method has greatly improved in accuracy and computational efficiency.

An Adaptive Grid Search Based Efficient Ensemble Model for Covid-19 Classification in Chest X-Ray Scans

Covid has resulted in millions of deaths worldwide, making it crucial to develop fast and safe diagnostic methods to control its spread. Chest X-Ray imaging can diagnose pulmonary diseases, including Covid. Most research studies have developed single convolution neural network models ignoring the advantage of combining different models. An ensemble model has higher predictive accuracy and reduces the generalization error of prediction. We employed an ensemble of Multi Deep Neural Networks models for Covid.19 classification in chest X-Ray scans using Multiclass classification (Covid, Pneumonia, and Normal). We improved the accuracy by identifying the best parameters using the sklean Grid search technique and implementing it with the Optimized Weight Average Ensemble Model, which allows multiple models to predict. Our ensemble model has achieved 95.26% accuracy in classifying the X-Ray images; it demonstrates potential in ensemble models for diagnosis using Radiography images.

Predicting Forex Currency Fluctuations Using a Novel Bio-Inspired Modular Neural Network

In the realm of foreign exchange (Forex) market predictions, Convolutional Neural Networks (CNNs) and Recurrent Neural Networks (RNNs) have been commonly employed. However, these models often exhibit instability due to vulnerability to data perturbations attributed to their monolithic architecture. Hence, this study proposes a novel neuroscience-informed modular network that harnesses closing prices and sentiments from Yahoo Finance and Twitter APIs. Compared to monolithic methods, the objective is to advance the effectiveness of predicting price fluctuations in Euro to British Pound Sterling (EUR/GBP). The proposed model offers a unique methodology based on a reinvigorated modular CNN, replacing pooling layers with orthogonal kernel initialisation RNNs coupled with Monte Carlo Dropout (MCoRNNMCD). It integrates two pivotal modules: a convolutional simple RNN and a convolutional Gated Recurrent Unit (GRU). These modules incorporate orthogonal kernel initialisation and Monte Carlo Dropout techniques to mitigate overfitting, assessing each module’s uncertainty. The synthesis of these parallel feature extraction modules culminates in a three-layer Artificial Neural Network (ANN) decision-making module. Established on objective metrics like the Mean Square Error (MSE), rigorous evaluation underscores the proposed MCoRNNMCD–ANN’s exceptional performance. MCoRNNMCD–ANN surpasses single CNNs, LSTMs, GRUs, and the state-of-the-art hybrid BiCuDNNLSTM, CLSTM, CNN–LSTM, and LSTM–GRU in predicting hourly EUR/GBP closing price fluctuations.

Unified CNN-LSTM for keyhole status prediction in PAW based on spatial-temporal features

Despite the high efficiency of keyhole plasma arc welding (K-PAW), it still has several deficiencies, such as narrow welding parameter ranges, easily disturbed welding process and instability of welding quality, etc. It is a significant prerequisite to predict the keyhole/penetration status accurately for maintaining the welding process stability and improving the welding quality. Most researchers have focused on visual inspection techniques and convolutional neural networks (CNN), establishing a mathematical model of the correlation between weld pool images and penetration status. While CNN could extract the features of single weld pool images, it is difficult to predict the evolution trend of the incoming moments. In this paper, a novel model based on CNN and LSTM (long-short term memory) was developed to extract both the spatial and temporal features of topside weld pool images, and consequently the complex keyhole behaviors were predicted and described. The comparative study is carried out on different models, i.e. the single CNN model, the single LSTM model, and the CNN-LSTM model that integrates both spatial features of single images and temporal features of sequence. The keyhole initialization and establishing period, as the typical and critical welding scenario in full-penetration welding, is investigate and compared, resulting in a predicted value that is close to reality. Furthermore, ahead prediction of keyhole status was adopted, maintaining over 80% accuracy even when predicting keyhole behaviors 2 s into the future. Consequently, the unified CNN-LSTM model effectively improves the prediction accuracy of keyhole/penetration status, promising for intelligent K-PAW technology.

Achieving wind power and photovoltaic power prediction: An intelligent prediction system based on a deep learning approach

Accurately predicting wind and photovoltaic power is one of the keys to improving the economy of wind-solar complementary power generation system, reducing scheduling costs and no-load losses, and ensuring grid stability. However, the natural properties of energy result in complex fluctuations in their corresponding power sequences, making accurate predictions difficult. Therefore, this paper proposes an intelligent prediction system that combines decomposition algorithms and deep learning for ultra-short-term prediction of wind and photovoltaic power. First, an improved decomposition algorithm is proposed, based on fuzzy entropy's property that its value increases with the increase of sequence uncertainty, particle swarm optimization (PSO) is used to search for the optimal parameter combinations of variational modal decomposition (VMD), so that it can automatically adjust the parameters for energy data with different characteristics to reduce the human error. Then, a convolutional neural network (CNN) architecture that balances operational efficiency and prediction performance is constructed, and the hyperparameters of the CNN are optimized using the wild horse optimization algorithm (WHO) to improve the stability and accuracy of the prediction model. In this paper, real data from wind power plants and photovoltaic power plants in China are used as experimental objects, and experiments are carried out in three aspects, namely, benchmark model selection, decomposition algorithm comparison and combined model comparison. The results show that selecting CNN as the benchmark model is a good choice; the improved VMD has better decomposition performance than other state-of-the-art decomposition algorithms. The system proposed in this paper is highly generalizable and adaptive, and its prediction performance and accuracy greatly outperform that of other comparative models, with prediction accuracies improved by 72% and 79%, respectively, compared to a single CNN model.

Assessment of the Object Detection Ability of Interproximal Caries on Primary Teeth in Periapical Radiographs Using Deep Learning Algorithms

The purpose of this study was to evaluate the performance of a model using You Only Look Once (YOLO) for object detection of proximal caries in periapical radiographs of children. A total of 2016 periapical radiographs in primary dentition were selected from the M6 database as a learning material group, of which 1143 were labeled as proximal caries by an experienced dentist using an annotation tool. After converting the annotations into a training dataset, YOLO was trained on the dataset using a single convolutional neural network (CNN) model. Accuracy, recall, specificity, precision, negative predictive value (NPV), F1-score, Precision-Recall curve, and AP (area under curve) were calculated for evaluation of the object detection model’s performance in the 187 test datasets. The results showed that the CNN-based object detection model performed well in detecting proximal caries, with a diagnostic accuracy of 0.95, a recall of 0.94, a specificity of 0.97, a precision of 0.82, a NPV of 0.96, and an F1-score of 0.81. The AP was 0.83. This model could be a valuable tool for dentists in detecting carious lesions in periapical radiographs.

Landfill Waste Segregation Using Transfer and Ensemble Machine Learning: A Convolutional Neural Network Approach

Waste disposal remains a challenge due to land availability, and environmental and health issues related to the main disposal method, landfilling. Combining computer vision (machine learning) and robotics to sort waste is a cost-effective solution for landfilling activities limitation. The objective of this study was to combine transfer and ensemble learning to process collected waste images and classify landfill waste into nine classes. Pretrained CNN models (Inception–ResNet-v2, EfficientNetb3, and DenseNet201) were used as base models to develop the ensemble network, and three other single CNN models (Models 1, 2, and 3). The single network performances were compared to the ensemble model. The waste dataset, initially grouped in two classes, was obtained from Kaggle, and reorganized into nine classes. Classes with a low number of data were improved by downloading additional images from Google search. The Ensemble Model showed the highest prediction precision (90%) compared to the precision of Models 1, 2, and 3, 86%, 87%, and 88%, respectively. All models had difficulties predicting overlapping classes, such as glass and plastics, and wood and paper/cardboard. The environmental costs for the Ensemble network, and Models 2 and 3, approximately 15 g CO2 equivalent per training, were lower than the 19.23 g CO2 equivalent per training for Model 1.

The effects of topological features on convolutional neural networks—an explanatory analysis via Grad-CAM

Topological data analysis (TDA) characterizes the global structure of data based on topological invariants such as persistent homology, whereas convolutional neural networks (CNNs) are capable of characterizing local features in the global structure of the data. In contrast, a combined model of TDA and CNN, a family of multimodal networks, simultaneously takes the image and the corresponding topological features as the input to the network for classification, thereby significantly improving the performance of a single CNN. This innovative approach has been recently successful in various applications. However, there is a lack of explanation regarding how and why topological signatures, when combined with a CNN, improve discriminative power. In this paper, we use persistent homology to compute topological features and subsequently demonstrate both qualitatively and quantitatively the effects of topological signatures on a CNN model, for which the Grad-CAM analysis of multimodal networks and topological inverse image map are proposed and appropriately utilized. For experimental validation, we utilize two famous datasets: the transient versus bogus image dataset and the HAM10000 dataset. Using Grad-CAM analysis of multimodal networks, we demonstrate that topological features enforce the image network of a CNN to focus more on significant and meaningful regions across images rather than task-irrelevant artifacts such as background noise and texture.

Images Classification Integrating Transformer with Convolutional Neural Network

Convolutional neural networks (CNN) are one of the most widely used deep learning methods in computer vision, which can effectively extract local spatial information from images, but lack global understanding and dependency modelling of image features. As a result, contextual information cannot be fully utilized by the network. For example, on coordinate modelling tasks (such as object detection, image generation, etc.), CNN may not be able to accurately locate or reconstruct the position and shape of objects. In contrast to traditional CNN models such as ResNet, Transformers rely on their global attention mechanism to capture long-distance dependencies between patches. The thesis presents an enhanced lightweight method which integrates Transformer with five convolutional neural layers. Model based on CNN and Transformer is tested on the two benchmark datasets MNIST and CIFAR-10. After a few epochs, the model is convergent and reaches high accuracy of 99.34% in MNIST and 92.04% in CIFAR-10. This model outperforms the single CNN and some state-of-the-art models in classifying both datasets, especially in distinguishing similar images like ‘6’ and ‘9’, ‘bird’ and ‘plane’. These results indicate the model's good robustness and generality.

Detecting glaucoma from fundus images using ensemble learning

Abstract Glaucomatous changes of the optic nerve head could be detected from fundus images. Focusing on optic nerve head appearance, and its difference from healthy images, altogether with the availability of plenty of such images in public fundus image databases, these images are ideal sources for artificial intelligence methods applications. In this work, we used ensemble learning methods and compared them with various single CNN models (VGG-16, ResNet-50, and MobileNet). The models were trained on images from REFUGE public dataset. The average voting ensemble method outperformed all mentioned models with 0.98 accuracy. In the AUC metric, the average voting ensemble method outperformed VGG-16 and MobileNet models, which had significantly weaker performance when used alone. The best results were observed using the ResNet-50 model. These results confirmed the significant potential of ensemble learning in enhancing the overall predictive performance in glaucomatous changes detection, but the overall performance could be negatively affected when models with weaker prediction performance are included.

Location Selection of Electric Vehicles Charging Stations Based on Analytical Hierarchy Process and Clustering Algorithm

In recent years, the charging infrastructure for electric vehicles (EVs) in China has been developing quickly, yet the charging station industry still faces problems such as unreasonable layout, difficulty in profitability and limited coverage. Based on this, this paper systematically compared the key factors of charging facility location and used the AHP method to measure the degree of influence of various factors at different levels on the location of charging stations of enterprises, so as to establish a scientific and reasonable evaluation model for the location of charging stations. Secondly, we selected 12 typical enterprises such as TELD NEW ENERGY Co., Ltd.from the dimensions of company scale and strategic objectives, and then classified them into state-owned enterprises, vehicle enterprises and specialized operating enterprises through k-means clustering. Finally it turns out that various types of enterprises are in the stage of differentiation in the investment and construction speed of charging station enterprises. In addition, this study proposes personalized siting solutions for different types of charging piles, and makes referenceable suggestions for the siting of charging stations.
 
 
 atasets MNIST and CIFAR-10. After a few epochs, the model is convergent and reaches high accuracy of 99.34% in MNIST and 92.04% in CIFAR-10. This model outperforms the single CNN and some state-of-the-art models in classifying both datasets, especially in distinguishing similar images like ‘6’ and ‘9’, ‘bird’ and ‘plane’. These results indicate the model's good robustness and generality.

FISRCN: a single small-sized image super-resolution convolutional neural network by using edge detection

FISRCN: a single small-sized image super-resolution convolutional neural network by using edge detection

Brain Computer Interface Based on Motor Imagery for Mechanical Arm Grasp Control

This paper puts forward a brain computer interface (BCI) system to realize the hand and wrist control using the ABB Mechanical Arm. This BCI system gathers four kinds of motor imaginary (MI) tasks (hand grasp, hand spread, wrist flexion and wrist extension) electroencephalogram (EEG) signals from 30 electrodes. It utilizes two fifth-order Butterworth Band-Pass Filter (BPF) with different bandwidths and normalization method to achieve the raw MI tasks EEG signals preprocessing. The main challenge of feature extraction is to extract enough representative features from MI tasks to classify them. This proposed BCI system extracts eleven kinds of features in time domain and time-frequency domain and uses mutual information method to reduce the large dimension of the extracted features. In addition, the BCI system applies a single convolutional layer Convolutional neural networks (CNN) with 30 filters to implement the quaternary classification of MI tasks. Compared with early researches, the classification accuracy of this BCI system is increased by about 35%. The actual mechanical arm grasping control experiments verifies that this BCI system has good adaptability.

Computer-Aided detection of tuberculosis from X-ray images using CNN and PatternNet classifier.

Tuberculosis (TB) is a highly infectious disease that mainly affects the human lungs. The gold standard for TB diagnosis is Xpert Mycobacterium tuberculosis/ resistance to rifampicin (MTB/RIF) testing. X-ray, a relatively inexpensive and widely used imaging modality, can be employed as an alternative for early diagnosis of the disease. Computer-aided techniques can be used to assist radiologists in interpreting X-ray images, which can improve the ease and accuracy of diagnosis. To develop a computer-aided technique for the diagnosis of TB from X-ray images using deep learning techniques. This research paper presents a novel approach for TB diagnosis from X-ray using deep learning methods. The proposed method uses an ensemble of two pre-trained neural networks, namely EfficientnetB0 and Densenet201, for feature extraction. The features extracted using two CNNs are expected to generate more accurate and representative features than a single CNN. A custom-built artificial neural network (ANN) called PatternNet with two hidden layers is utilized to classify the extracted features. The effectiveness of the proposed method was assessed on two publicly accessible datasets, namely the Montgomery and Shenzhen datasets. The Montgomery dataset comprises 138 X-ray images, while the Shenzhen dataset has 662 X-ray images. The method was further evaluated after combining both datasets. The method performed exceptionally well on all three datasets, achieving high Area Under the Curve (AUC) scores of 0.9978, 0.9836, and 0.9914, respectively, using a 10-fold cross-validation technique. The experiments performed in this study prove the effectiveness of features extracted using EfficientnetB0 and Densenet201 in combination with PatternNet classifier in the diagnosis of tuberculosis from X-ray images.

Data Fusion for Cross-Domain Real-Time Object Detection on the Edge

We investigate an edge-computing scenario for robot control, where two similar neural networks are running on one computational node. We test the feasibility of using a single object-detection model (YOLOv5) with the benefit of reduced computational resources against the potentially more accurate independent and specialized models. Our results show that using one single convolutional neural network (for object detection and hand-gesture classification) instead of two separate ones can reduce resource usage by almost 50%. For many classes, we observed an increase in accuracy when using the model trained with more labels. For small datasets (a few hundred instances per label), we found that it is advisable to add labels with many instances from another dataset to increase detection accuracy.

EfficientNetV2-based dynamic gesture recognition using transformed scalogram from triaxial acceleration signal

Abstract In this paper, a dynamic gesture recognition system is proposed using triaxial acceleration signal and image-based deep neural network. With our dexterous glove device, 1D acceleration signal can be measured from each finger and decomposed to time-divided frequency components via wavelet transformation, which is known as scalogram as image-like format. To feed-forward the scalogram with single 2D, convolutional neural networks allows the gesture having temporality to be easily recognized without any complex system such as RNN, LSTM, or spatio-temporal feature as 3D CNN, etc. To classify the image with general input dimension of image RGB channels, we numerically reconstruct fifteen scalograms into one RGB image with various representation methods. In experiments, we employ the off-the-shelf model, EfficientNetV2 small-to-large model as an image classification model with fine-tuning. To evaluate our system, we bulid our custom bicycle hand signals as dynamic gesture dataset under our transformation system, and then qualitatively compare the reconstruction method with matrix representation methods. In addition, we use other signal transformation tools such as the fast Fourier transform and short-time Fourier transform and then explain the advantages of scalogram classification in the terms of time-frequency resolution trade-off issue.

A Leading but Simple Classification Method for Remote Sensing Images

Recently, researchers have proposed a lot of deep convolutional neural network (CNN) approaches with obvious flaws to tackle the difficult semantic classification (SC) task of remote sensing images (RSI). In this paper, the author proposes a simple method that aims to provide a leading but efficient solution by using a lightweight EfficientNet-B0. First, this paper concluded the drawbacks with an analysis of mathematical theory and then proposed a qualitative conclusion on the previous methods’ theoretical performance based on theoretical derivation and experiments. Following that, the paper designs a novel method named LS-EfficientNet, consisting only of a single CNN and a concise training algorithm called SC-CNN. Far different from previous complex and hardware-extensive ones, the proposed method mainly focuses on tackling the long-neglected problems, including overfitting, data distribution shift by DA, improper use of training tricks, and other incorrect operations on a pre-trained CNN. Compared to previous studies, the proposed method is easy to reproduce because all the models, training tricks, and hyperparameter settings are open-sourced. Extensive experiments on two benchmark datasets show that the proposed method can easily surpass all the previous state-of-the-art ones, with an outstanding accuracy lead of 0.5% to 1.2% and a remarkable parameter decrease of 78% if compared to the best prior one in 2022. In addition, ablation test results also prove that the proposed effective combination of training tricks, including OLS and CutMix, can clearly boost a CNN's performance for RSI-SC, with an increase in accuracy of 1.0%. All the results reveal that a single lightweight CNN can well tackle the routine task of classifying RSI.

Multi-step CNN forecasting for COVID-19 multivariate time-series

The new coronavirus (COVID-19) has spread to over 200 countries, with over 36 million confirmed cases as of October 10, 2020. As a result, numerous machine learning models capable of forecasting the epidemic worldwide have been produced. This paper reviews and summarizes the most relevant machine learning forecasting models for COVID-19. The dataset is derived from the world health organization (WHO) COVID-19 dashboard, and it contains official daily counts of COVID-19 cases, fatalities, and vaccination use reported by countries, territories, and regions. We propose various convolutional neural network (CNN) based models such as CNN, single exponential smoothing CNN (S-CNN), moving average CNN (MA-CNN), smoothed moving average CNN (SMA-CNN), and moving average smoothed CNN (MAS-CNN). Here, MAPE and MSE are used to assess the suggested models. MAPE is frequently used to compare accuracy across time series with different scales. MSE, the model must strive for a total forecast equal to the entire demand. That is, optimizing MSE seeks to create a forecast that is right on average and so unbiased. The final result shows that SMA-CNN outperformed its baselines in both MAPE and MSE. The main contribution of this novel forecasting approach is a more accurate result as a base of the strategy of preventing COVID-19 spreads.

DBCT-Net:A dual branch hybrid CNN-transformer network for remote sensing image fusion

Remote sensing image fusion aims at fusing high spatial resolution single-band panchromatic (PAN) image with spectrally informative multispectral (MS) image to generate panchromatic sharpened image with high resolution and color information, it is also called pansharpening. Most of the proposed single convolutional neural network (CNN) or transformer-based pansharpening methods own several problems, such as inability to acquire long-range features or difficult to train, resulting the loss of spatial details and colors. In addition, the computational complexity of transformer cannot be ignored. In this work, we propose a dual-branch hybrid CNN-Transformer network (DBCT-Net) that utilizes the local specificity of CNN and models the global dependencies by transformer. First, a multi-branch dense connected block (MDCB-4) network is designed to obtain spectral and textural information in MS and PAN images, respectively. Next, an encoder–decoder transformer based on the self-attention and co-attention modules is able to inject the missing local and global information, which can further enhance the results. It is worth noting that an inverted multi-head transposed attention (IMTA) is applied here to build attention maps from feature dimensions, which greatly reduces the computation time. Finally, an image reconstruction module is employed to effectively fuse the acquired texture and spectral features. Furthermore, to generate visually better pansharpened images, we propose a combined loss function that includes a focal frequency loss. Extensive experiments on WorldView II (WV2), GF-2,and QuickBird (QB) datasets show that DBCT-Net can perform better in spatial preservation and spectral feature recovery.