Convolutional Neural Network CNN Research Articles

Course genres classification of music e-learning platform based on deep learning big data intelligent processing algorithm

At present, the growth trend of digital music information is showing an exponential form. Various songs can be classified according to their unique genres and styles, forming various “genres”. So, how to effectively manage the massive amount of music data has become a problem that needs to be solved to promote the balanced development of different types and genres of music. On this basis, this article adopts deep learning technology to study the classification methods of music genres oriented towards big data. Firstly, this article studies LSTM based on long and short term memory networks and compares its advantages and disadvantages with convolutional neural network CNN. Secondly, this article calculates the drift patterns of big data intelligent processing algorithms and conducts relevant performance tests on big data intelligent processing algorithms to verify the effectiveness of the proposed algorithms. Finally, this article addresses the problems in current music genre classification research by introducing Mel filters into the extraction of sound features to further improve the efficiency and speed of the model. And deep learning technology is adopted to construct a music genre classification model based on big data intelligent analysis, further promoting the effective improvement of music genre classification effectiveness.

The dependence of the effectiveness of neural networks for recognizing human voice on language

This study examines the effectiveness of neural network architectures (multilayer perceptron MLP, convolutional neural network CNN, recurrent neural network RNN) for human voice recognition, with an emphasis on the Kazakh language. Problems related to language, the difference between speakers, and the influence of network architecture on recognition accuracy are considered. The methodology includes extensive training and testing, studying the accuracy of recognition in different languages, and different sets of data on speakers. Using a comparative analysis, this study evaluates the performance of three architectures trained exclusively in the Kazakh language. The testing included statements in Kazakhs and other languages, while the number of speakers varied to assess its impact on recognition accuracy. During the study, the results showed that CNN neural networks are more effective in recognizing human voice than RNN and MLP. Also, it was found that the CNN has a higher accuracy in recognizing the human voice in the Kazakh language, both for a small and for a large number of announcers. For example, for 20 speakers, the recognition error in Russian was 21.86 %, whereas in Kazakhs it was 10.6 %. A similar trend was observed for 80 speakers: 16.2 % Russians and 8.3 % Kazakhs. It can also be argued that learning one language does not guarantee high recognition accuracy in other languages. Therefore, the accuracy of human voice recognition by neural networks depends significantly on the language in which training is conducted. In addition, this study highlights the importance of different sets of speaker data to achieve optimal results. This knowledge is crucial for advancing the development of reliable human voice recognition systems that can accurately identify different human voices in different language contexts

Integrated Multi-Head Self-Attention Transformer model for electricity demand prediction incorporating local climate variables

This paper develops a trustworthy deep learning model that considers electricity demand (G) and local climate conditions. The model utilises Multi-Head Self-Attention Transformer (TNET) to capture critical information from G, to attain reliable predictions with local climate (rainfall, radiation, humidity, evaporation, and maximum and minimum temperatures) data from Energex substations in Queensland, Australia. The TNET model is then evaluated with deep learning models (Long-Short Term Memory LSTM, Bidirectional LSTM BILSTM, Gated Recurrent Unit GRU, Convolutional Neural Networks CNN, and Deep Neural Network DNN) based on robust model assessment metrics. The Kernel Density Estimation method is used to generate the prediction interval (PI) of electricity demand forecasts and derive probability metrics and results to show the developed TNET model is accurate for all the substations. The study concludes that the proposed TNET model is a reliable electricity demand predictive tool that has high accuracy and low predictive errors and could be employed as a stratagem by demand modellers and energy policy-makers who wish to incorporate climatic factors into electricity demand patterns and develop national energy market insights and analysis systems.

Sentiment Analysis of Online New Energy Vehicle Reviews

Massive online reviews of new energy vehicles in China are deemed crucial by companies, as they offer valuable insights into user demands and perceptions. An effective analysis enables companies to swiftly adapt and enhance their products while upholding a positive public image. Nonetheless, the sentiment analysis of online car reviews can pose challenges due to factors such as incomplete context, abrupt information bursts, and specialized vocabulary. In this paper, an enhanced hybrid model is introduced, combining Enhanced Representation through kNowledge IntEgration (ERNIE) and a deep (Convolutional Neural Network) CNN, to tackle these challenges. The model utilizes fine-tuned ERNIE for feature extraction from preprocessed review datasets, generating word vectors that encompass comprehensive semantic information. The deep CNN component captures local features from the text, thereby capturing semantic nuances at multiple levels. To address sudden shifts in public sentiment, a channel attention mechanism is employed to amplify the significance of crucial information within the reviews, facilitating comment relationship classification and sentiment prediction. The experimental results demonstrate the efficacy of the proposed model, achieving an impressive accuracy rate of 97.39% on the test set and significantly outperforming other models.

Feature Deep Learning Extraction Approach for Object Detection in Self-Driving Cars

Self-driving cars are a fundamental research subject in recent years; the ultimate goal is to completely exchange the human driver with automated systems. On the other hand, deep learning techniques have revealed performance and effectiveness in several areas. The strength of self-driving cars has been deeply investigated in many areas including object detection, localization as well, and activity recognition. This paper provides an approach to deep learning; which combines the benefits of both convolutional neural network CNN together with Dense technique. This approach learns based on features extracted from the feature extraction technique which is linear discriminant analysis LDA combined with feature expansion techniques namely: standard deviation, min, max, mod, variance and mean. The presented approach has proven its success in both testing and training data and achieving 100% accuracy in both terms.

Saca-FI: A microarchitecture-level fault injection framework for reliability analysis of systolic array based CNN accelerator

As convolutional neural network CNN accelerators are being adopted in emerging safety-critical areas, their reliability becomes prominent. The systolic array is widely used as the major processing structure in CNN accelerators, so its reliability features need to be comprehensively explored. In this work, we develop a microarchitecture-level fault injection framework, saca-FI, to analyze the reliability of systolic array based CNN accelerators. It is able to investigate various vulnerability sources (e.g., single-bit flips, multi-bit flips, stuck-at faults, multiple faults) and at different levels (e.g., model layer and storage locations). Based on saca-FI, we evaluate the resilience of several commonly used networks, and observe several important architecture-level reliability characteristics. We then propose two opportunistic protections to demonstrate the usage of saca-FI in designing energy-efficient error protection mechanisms.

Efficient Hybrid Algorithm for Human Action Recognition

Recently, researchers have sought to find the ideal way to recognize human actions through video using artificial intelligence due to the multiplicity of applications that rely on it in many fields. In general, the methods have been divided into traditional methods and deep learning methods, which have provided a qualitative leap in the field of computer vision. Convolutional neural network CNN and recurrent neural network RNN are the most popular algorithms used with images and video. The researchers combined the two algorithms to search for the best results in a lot of research. In an attempt to obtain improved results in motion recognition through video, we present in this paper a combined algorithm, which is divided into two main parts, CNN and RNN. In the first part there is a preprocessing stage to make the video frame suitable for the input of both CNN networks which consist of a fusion of Inception-ResNet-V2 and GoogleNet to obtain activations, with the previously trained wights in Inception-ResNet-V2 and GoogleNet and then passed to a deep Gated Recurrent Units (GRU) connected to a fully connected SoftMax layer to recognize and distinguish the human action in the video. The results show that the proposed algorithm gives better accuracy of 97.97% with the UCF101 dataset and 73.12% in the hdmb51 data set compared to those present in the related literature.

ImECGnet: Cardiovascular Disease Classification from Image-Based ECG Data Using a Multibranch Convolutional Neural Network

Reliable Cardiovascular Disease (CVD) classification performed by a smart system can assist medical doctors in recognizing heart illnesses in patients more efficiently and effectively. Electrocardiogram (ECG) signals are an important diagnostic tool as they are already available early in the patients’ health diagnosis process and contain valuable indicators for various CVDs. Most ECG processing methods represent ECG data as a time series, often as a matrix with each row containing the measurements of a sensor lead; and/or the transforms of such time series like wavelet power spectrums. While methods processing such time-series data have been shown to work well in benchmarks, they are still highly dependent on factors like input noise and sequence length, and cannot always correlate lead data from different sensors well. In this paper, we propose to represent ECG signals incorporating all lead data plotted as a single image, an approach not yet explored by literature. We will show that such an image representation combined with our newly proposed convolutional neural network specifically designed for CVD classification can overcome the aforementioned shortcomings. The proposed (Convolutional Neural Network) CNN is designed to extract features representing both the proportional relationships of different leads to each other and the characteristics of each lead separately. Empirical validation on the publicly available PTB, MIT-BIH, and St.-Petersburg benchmark databases shows that the proposed method outperforms time seriesbased state-of-the-art approaches, yielding classification accuracy of 97.91%, 99.62%, and 98.70%, respectively.

Prediction of Effective Elastic and Thermal Properties of Heterogeneous Materials Using Convolutional Neural Networks

The aim of this study is to develop a new method to predict the effective elastic and thermal behavior of heterogeneous materials using Convolutional Neural Networks CNN. This work consists first of all in building a large database containing microstructures of two phases of heterogeneous material with different shapes (circular, elliptical, square, rectangular), volume fractions of the inclusion (20%, 25%, 30%), and different contrasts between the two phases in term of Young modulus and also thermal conductivity. The contrast expresses the degree of heterogeneity in the heterogeneous material, when the value of C is quite important (C >> 1) or quite low (C << 1), it means that the material is extremely heterogeneous, while C= 1, the material becomes totally homogeneous. In the case of elastic properties, the contrast is expressed as the ratio between Young’s modulus of the inclusion and that of the matrix (C = EiEm), while for thermal properties, this ratio is expressed as a function of the thermal conductivity of both phases (C = λiλm). In our work, the model will be tested on two values of contrast (10 and 100). These microstructures will be used to estimate the elastic and thermal behavior by calculating the effective bulk, shear, and thermal conductivity values using a finite element method. The collected databases will be trained and tested on a deep learning model composed of a first convolutional network capable of extracting features and a second fully connected network that allows, through these parameters, the adjustment of the error between the found output and the expected one. The model was verified using a Mean Absolute Percentage Error (MAPE) loss function. The prediction results were excellent, with a prediction score between 92% and 98%, which justifies the good choice of the model parameters.

Identification of Acral Melanoma using Fuzzy Algorithm compared with Convolutional Neural Network for improved accuracy using Dermoscopic Images

Aim: Acral melanoma detection using a fuzzy method vs. convolutional neural network CNN utilizing dermoscopic pictures. Materials and Methods: To assess and compare the acral melanoma detection, the researchers used the fuzzy algorithm and the Convolutional neural network algorithm. The overall sample size is 40 due to the utilization of 20 samples acral melanoma is identified by evaluating the effectiveness with pretest power of 80% (G-power) α=0.05, confidence level 95%. Result: The proposed fuzzy algorithm helps in increasing the higher accuracy compared to Convolutional Neural Network CNN with improved accuracy of the fuzzy algorithm algorithm is 80% and the CNN algorithm is 75 %. The accurate rate is 80with data features found in fuzzy algorithms and each algorithm has different precision. Conclusion: This study shows a higher accuracy for the fuzzy algorithm higher significant value than the convolutional neural network.

Identification of Acral Melanoma using Genetic Algorithms Compared with Convolutional Neural Network using Dermoscopic Images

Aim: Identification of acral melanoma using genetic algorithm compared with convolutional neural network CNN using dermoscopic images. Materials and Methods: The study was conducted using the genetic algorithm and convolutional neural network algorithm to analyze and compare the acral melanoma detection. The number of samples used is 20, total sample size is 40. Acral melanoma is identified by evaluating the effectiveness with pre-test power of 80% (G-power), α=0.05, confidence interval 95%. Result: The proposed genetic algorithm helps in increasing the higher accuracy compared to convolutional neural networks with improved accuracy of the genetic algorithm algorithm is 96 % and the convolutional neural network algorithm is 95%. The accurate rate is 80 with the data features found in the genetic algorithm algorithm. Precision is different in each algorithm. Conclusion: This study shows a higher accuracy for the genetic algorithm when compared with convolutional neural networks.

The Impact of Renewable Electricity Output on Sustainability in the Context of Circular Economy: A Global Perspective

In this article, we investigate the impact of “Renewable Electricity Output” on the green economy in the context of the circular economy for 193 countries in the period 2011–2020. We use data from the World Bank ESG framework. We perform Panel Data with Fixed Effects, Panel Data with Random Effects, Weighted Last Squares-WLS, and Pooled Ordinary Least Squares-OLS. Our results show that Renewable Electricity Output is positively associated, among others, with “Adjusted Savings-Net Forest Depletion” and “Renewable Energy Consumption” and negatively associated, among others, with “CO2 Emission” and “Cooling Degree Days”. Furthermore, we perform a cluster analysis implementing the k-Means algorithm optimized with the Elbow Method and we find the presence of four clusters. In adjunct, we confront seven different machine learning algorithms to predict the future level of “Renewable Electricity Output”. Our results show that Linear Regression is the best algorithm and that the future value of renewable electricity output is predicted to growth on average at a rate of 0.83% for the selected countries. Furthermore, we improve the machine learning analysis with a Deep Learning approach using Convolutional Neural Network-CNN but the algorithm is not appropriate for the analyzed dataset. Less complex machine learning algorithms show better statistical results.

Mostar Index of Neural Networks and Applications

Neural networks are mathematical models that use learning algorithms inspired by the brain to store information. These networks have been extensively used to operate cognitive functions, robustness, fault tolerance, flexibility, collective computation, and ability to deal with fuzzy and inconsistent information. In recent years, probabilistic neural network PNN and convolutional neural network CNN have been employed in optimization problems such as input normalization and decision making tasks. The flow of data in PNN and CNN will be optimized if its peripherality is minimized. Topological index (TI) is a graph invariant initially introduced to study physiochemical properties of chemical networks. They are further applicable to study certain mathematical aspects of a network such as irregularity, centrality and peripherality. Recently, the Mostar index and edge Mostar index have been introduced to study peripherality of various networks. In the current work, Mostar indices for PNN and CNN have been computed and their application in context of data flow problem together with graphical analysis is presented.

Feature Extraction for Image Analysis and Detection using Machine Learning Techniques

Feature extraction is the most vital step in image classification to produce high-quality and good content images for further analysis, image detection, segmentation, and object recognition. Using machine learning algorithms, profound learning like convolutional neural network CNN became necessary to train, classify, and recognize images and objects like humans. Combined feature extraction and machine learning classification to locate and identify objects on images can then be an input of automatic recognition systems ATR such as surveillance systems CCTV, to enhance these systems and reduce time and effort for object detection and recognition in images based on digital image processing techniques especially image segmentation that differentiate from computer vision approach. This article will use machine learning and deep learning algorithms to facilitate and achieve the study's objectives.

An Investigation on Disease Diagnosis and Prediction by Using Modified K-Mean clustering and Combined CNN and ELM Classification Techniques

Data analysis is important for managing a lot of knowledge in the healthcare industry. The older medical study favored prediction over processing and assimilating a massive volume of hospital data. The precise research of health data becomes advantageous for early disease identification and patient treatment as a result of the tremendous knowledge expansion in the biological and healthcare fields. But when there are gaps in the medical data, the accuracy suffers. The use of K-means algorithm is modest and efficient to perform. It is appropriate for processing vast quantities of continuous, high-dimensional numerical data. However, the number of clusters in the given dataset must be predetermined for this technique, and choosing the right K is frequently challenging. The cluster centers chosen in the first phase have an impact on the clustering results as well. To overcome this drawback in k-means to modify the initialization and centroid steps in classification technique with combining (Convolutional neural network) CNN and ELM (extreme learning machine) technique is used. To increase this work, disease risk prediction using repository dataset is proposed. We use different types of machine learning algorithm for predicting disease using structured data. The prediction accuracy of using proposed hybrid model is 99.8% which is more than SVM (support vector machine), KNN (k-nearest neighbors), AB (AdaBoost algorithm) and CKN-CNN (consensus K-nearest neighbor algorithm and convolution neural network).

STUDY OF NEURAL NETWORKS FOR FORECASTING THE VALUE OF COMPANY SHARES IN AN UNSTABLE ECONOMY

These studies deal with analysis and selection of neural networks with various architectures and hybrid models, which include neural networks, to predict the market value of shares in the stock market of a country that is in the process of unstable development. Analysis and forecasting of such stock markets cannot be carried out using classical methods. The relevance of the research topic is due to the need to develop software systems that implement algorithmic support for predicting the market value of shares in Ukraine. The introduction of such software systems in the circuit of investment decisionmaking in companies that are interested in increasing the information transparency of the Ukrainian stock market will improve the forecasts of the market value of shares. This, in turn, will help improve the investment climate and ensure the growth of investment in the Ukrainian economy. The analysis of the results of existing studies on the use of neural networks and other methods of computational intelligence for modeling the behavior of stock market participants and market forecasting has been carried out. The article presents the results of a study for the using of neural networks with various architectures for predicting the market value of shares in the stock markets of Ukraine. Four shares of the Ukrainian Stock Exchange were chosen for forecasting: Centrenergo (CEEN); Ukrtelecom (UTLM); Kriukivs’kyi Vahonobudivnyi Zavod PAT (KVBZ); Raiffeisen Bank Aval (BAVL). The following models were chosen for the experimental study: long short-term memory LSTM; convolutional neural network CNN; a hybrid model combining two neural networks CNN and LSTM; a hybrid model consisting of a variational mode decomposition algorithm and a long-term memory neural network (VMD-LSTM); hybrid VMD-CNN-LSTM deep learning model based on variational mode (VMD) and two neural networks. Estimates of forecast quality based on various metrics were calculated. It is concluded that the use of the hybrid model VMD-CNN-LSTM gives the minimum error in predicting the market value of the shares of Ukrainian enterprises. It is also advisable to use the VMD-LSTM model to predict the stock exchanges of countries with an unstable economy.

Based on Wavelet Threshold Denoising-LDA and Bilstm Aircraft Engine Life Prediction

Turbine engines operate in extreme environments and require regular maintenance and replacement. It leads to a lot of human and financial investment, and once the aircraft fails during the navigation process, it will cause irreparable losses. Therefore, it is very necessary to propose an engine life prediction model, which can intervene manually before the engine fails, thereby reducing the probability of an air accident. In order to solve the problem, this paper proposes a wavelet threshold denoising-LDA and Bilstm aero-engine life prediction method, and uses the C-MAPSS data set for simulation prediction. Based on this data set, this paper firstly carried out wavelet threshold denoising to the data set to remove noise, and then used LDA to reduce the dimension to screen out the main features. The processed features are input into the bidirectional LSTM network for prediction. By training the above preprocessed data, BIlstm has better adaptability than the convolutional neural network CNN and LSTM, and the prediction error is greatly reduced.

Distillation-Guided Residual Learning for Binary Convolutional Neural Networks.

It is challenging to bridge the performance gap between binary convolutional neural network (BCNN) and floating-point CNN (FCNN). This performance gap is mainly caused by the inferior modeling capability and training strategy of BCNN, which leads to substantial residuals in intermediate feature maps between BCNN and FCNN. To minimize the performance gap, we enforce BCNN to produce similar intermediate feature maps with the ones of FCNN. This intuition leads to a more effective training strategy for BCNN, i.e., optimizing each binary convolutional block with blockwise distillation loss derived from FCNN. The goal of minimizing the residuals in intermediate feature maps also motivates us to update the binary convolutional block architecture to facilitate the optimization of blockwise distillation loss. Specifically, a lightweight shortcut branch is inserted into each binary convolutional block to complement residuals at each block. Benefited from its squeeze-and-interaction (SI) structure, this shortcut branch introduces a fraction of parameters, e.g., less than 10% overheads, but effectively boosts the modeling capability of binary convolution blocks in BCNN. Extensive experiments on ImageNet demonstrate the superior performance of our method in both classification efficiency and accuracy, e.g., BCNN trained with our methods achieves the accuracy of 60.45% on ImageNet, better than many state-of-the-art ones.

Integration of Object-Based Image Analysis and Convolutional Neural Network for the Classification of High-Resolution Satellite Image: A Comparative Assessment

During the past decade, deep learning-based classification methods (e.g., convolutional neural networks—CNN) have demonstrated great success in a variety of vision tasks, including satellite image classification. Deep learning methods, on the other hand, do not preserve the precise edges of the targets of interest and do not extract geometric features such as shape and area. Previous research has attempted to address such issues by combining deep learning with methods such as object-based image analysis (OBIA). Nonetheless, the question of how to integrate those methods into a single framework in such a way that the benefits of each method complement each other remains. To that end, this study compared four integration frameworks in terms of accuracy, namely OBIA artificial neural network (OBIA ANN), feature fusion, decision fusion, and patch filtering, according to the results. Patch filtering achieved 0.917 OA, whereas decision fusion and feature fusion achieved 0.862 OA and 0.860 OA, respectively. The integration of CNN and OBIA can improve classification accuracy; however, the integration framework plays a significant role in this. Future research should focus on optimizing the existing CNN and OBIA frameworks in terms of architecture, as well as investigate how CNN models should use OBIA outputs for feature extraction and classification of remotely sensed images.

Traffic Sign Recognition and Classification Using Deep Convolution Neural Network

Traffic signs classification is that the process of identifying which class a traffic sign belongs to. There are several differing kinds of traffic signs like speed limits, no entry, turn left or right, etc. A neural network model based on deep learning is utilized to explore the traffic sign recognition (TSR) and expand the application of deep intelligent learning technology in the field of virtual reality (VR) image recognition, thereby assessing the road traffic safety risks and promoting the construction of intelligent transportation networks Firstly, a dual-path deep CNN (TDCNN) TSR model is constructed supported CNN, and recognition accuracy is calculated to research the training results of the model. Secondly, we assess the road traffic safety risks, and also the prediction and evaluation effects. Finally, the changes in safety risks of road traffic accidents are analyzed based on the two key influencing factors of the number of road intersections and the speed of vehicles traveling. The results show that the learning rate of the network model and the number of hidden neurons in the fully-connected layer directly affect the training results, and there are differences in the choices between the early and late training periods. It instantly assists drivers or automatic driving systems in detecting and recognizing traffic signs effectively. Key Words: Deep Convolutional Neural Network CNN, TensorFlow, Image Processing, ReLU Activation Function.