Use Of Convolutional Neural Networks Research Articles

Convolutional Neural Network Approach for Mobile Banking Fraudulent Transaction to Detect Financial Frauds

In the real world, the identification of financial fraud in compliance with IoT criteria is highly effective since financial fraud causes financial damage. Several forms of financial fraud are likely, but unauthorized usage of mobile payment by credit card no. or certificate no.is the most common scenario. Detection of financial crime is a growing environment in which the victims will keep ahead. However, intelligent fraud detection facets remain scientifically unsupported. Deep learning (DL)arises from the idea of a multi-type representation of the human brain that incorporates basic characteristics at the low level or high-level abstractions.Financial fraud was a big issue as forgers discovered new methods of stealing currency. Therefore, adaptive methods of identification of fraud against forgers are required. Thanks to their versatile nature to detect emergent financial transaction fraud, deep learning approaches were enticing candidates. In this article, we suggest an in-depth learning approach for adapting financial fraud through the use of convolution neural networks (CNN). With the fraudulent transactions dataset, we tested our model experimentally. The results of the analysis show which our methods detect transactional fraud appropriately.

Interpretation of Convolutional Neural Networks for Acid Sulfate Soil Classification

Convolutional neural networks (CNNs) have been originally used for computer vision tasks, such as image classification. While several digital soil mapping studies have been assessing these deep learning algorithms for the prediction of soil properties, their potential for soil classification has not been explored yet. Moreover, the use of deep learning and neural networks in general has often raised concerns because of their presumed low interpretability (i.e., the black box pitfall). However, a recent and fast-developing sub-field of Artificial Intelligence (AI) called explainable AI (XAI) aims to clarify complex models such as CNNs in a systematic and interpretable manner. For example, it is possible to apply model-agnostic interpretation methods to extract interpretations from any machine learning model. In particular, SHAP (SHapley Additive exPlanations) is a method to explain individual predictions: SHAP values represent the contribution of a covariate to the final model predictions. The present study aimed at, first, evaluating the use of CNNs for the classification of potential acid sulfate soils located in the wetland areas of Jutland, Denmark (c. 6,500 km2), and second and most importantly, applying a model-agnostic interpretation method on the resulting CNN model. About 5,900 soil observations and 14 environmental covariates, including a digital elevation model and derived terrain attributes, were utilized as input data. The selected CNN model yielded slightly higher prediction accuracy than the random forest models which were using original or scaled covariates. These results can be explained by the use of a common variable selection method, namely recursive feature elimination, which was based on random forest and thus optimized the selection for this method. Notably, the SHAP method results enabled to clarify the CNN model predictions, in particular through the spatial interpretation of the most important covariates, which constitutes a crucial development for digital soil mapping.

Application of convolutional neural networks for prediction of disinfection by-products

Fluorescence spectroscopy can provide high-level chemical characterization and quantification that is suitable for use in online process monitoring and control. However, the high-dimensionality of excitation–emission matrices and superposition of underlying signals is a major challenge to implementation. Herein the use of Convolutional Neural Networks (CNNs) is investigated to interpret fluorescence spectra and predict the formation of disinfection by-products during drinking water treatment. Using deep CNNs, mean absolute prediction error on a test set of data for total trihalomethanes, total haloacetic acids, and the major individual species were all < 6 µg/L and represent a significant difference improved by 39–62% compared to multi-layer perceptron type networks. Heat maps that identify spectral areas of importance for prediction showed unique humic-like and protein-like regions for individual disinfection by-product species that can be used to validate models and provide insight into precursor characteristics. The use of fluorescence spectroscopy coupled with deep CNNs shows promise to be used for rapid estimation of DBP formation potentials without the need for extensive data pre-processing or dimensionality reduction. Knowledge of DBP formation potentials in near real-time can enable tighter treatment controls and management efforts to minimize the exposure of the public to DBPs.

Using convolutional neural networks to classify melt pools in a pulsed selective laser melting process

In a selective laser melting (SLM) process, the fluctuation of as-produced part quality remains a problem that hinders its broader industrial applications. In situ monitoring that identifies the variations of melt pool characteristics is promising to improve part quality and control the SLM process. So far, such monitoring approaches have been impeded by the limitation of image processing methods that cannot classify large volumes of melt pool images during a short time. Here we demonstrate that convolutional neural networks (CNN) could be an effective, efficient, and reliable solution to improve the SLM process. In this study, we employ the images captured at five input laser energy levels to train and test CNN models. Compared with a statistical method, the use of CNN has improved the classification of melt pools to the highest achieved accuracy of 96.6%, which is resulted from the ability of utilizing complex image features from the melt pool region. Thus, the CNN is a reliable method with the potential to monitor the influence of other process parameters on melt pool characteristics, potentially leading to the improvement of production in SLM.

Predicting the sparks occurrence in electrochemical discharge machining by machine learning using convolutional neural networks

This study investigates the use of convolutional neural networks (CNNs) to define sparks from the high-speed video feed of the electrochemical discharge machining (ECDM) process. The visual data is used to monitor the spark activity in the electrolyte. The recognition of the sparks in optical data can potentially improve the prediction of material removal in ECDM since the majority of machining is caused by the sparking. The massive dataset size is a challenge to study the optical data frame by frame. The CNN model in this study generated a time series for the presence of sparks based on the image feed in sequential order. The CNN based machine learning model in this study is found to be more consistent than the manual labeling of the images. This model is used to analyze the image data and predict the presence of the sparks over 95% accuracy.

Artificial Intelligence in Diagnostic Radiology: Where Do We Stand, Challenges, and Opportunities.

Artificial intelligence (AI) is the most revolutionizing development in the health care industry in the current decade, with diagnostic imaging having the greatest share in such development. Machine learning and deep learning (DL) are subclasses of AI that show breakthrough performance in image analysis. They have become the state of the art in the field of image classification and recognition. Machine learning deals with the extraction of the important characteristic features from images, whereas DL uses neural networks to solve such problems with better performance. In this review, we discuss the current applications of machine learning and DL in the field of diagnostic radiology.Deep learning applications can be divided into medical imaging analysis and applications beyond analysis. In the field of medical imaging analysis, deep convolutional neural networks are used for image classification, lesion detection, and segmentation. Also used are recurrent neural networks when extracting information from electronic medical records and to augment the use of convolutional neural networks in the field of image classification. Generative adversarial networks have been explicitly used in generating high-resolution computed tomography and magnetic resonance images and to map computed tomography images from the corresponding magnetic resonance imaging. Beyond image analysis, DL can be used for quality control, workflow organization, and reporting.In this article, we review the most current AI models used in medical imaging research, providing a brief explanation of the various models described in the literature within the past 5 years. Emphasis is placed on the various DL models, as they are the most state-of-art in imaging analysis.

Uncertainty Estimation via Monte Carlo Dropout in CNN-Based mmWave MIMO Localization

Recently, there has been much interest in the use of convolutional neural networks (CNN) for mobile user localization in massive multiple-input multiple-output (MIMO) systems operating at millimeter wave (mmWave) frequencies. However, current CNN-based approaches cannot predict the confidence interval bounds for the localization accuracy. While the Bayesian neural network (BNN) method can be employed to estimate the model uncertainty, it entails a high computational cost. In this letter, the Monte Carlo (MC) dropout based method is proposed as a low-complexity approximation to BNN inference for capturing the uncertainty in a CNN-based mmWave MIMO outdoor localization system, without sacrificing accuracy. The proposed method is evaluated by means of simulations using a ray-tracing model of urban propagation at 28GHz. Results show that the localization uncertainty region can be properly determined and that their shape depends on the maximum power received at the user.

Spectral Feature Fusion Networks With Dual Attention for Hyperspectral Image Classification

Recent progress in spectral classification is largely attributed to the use of convolutional neural networks (CNNs). While a variety of successful architectures have been proposed, they all extract spectral features from various portions of adjacent spectral bands. In this article, we take a different approach and develop a deep spectral feature fusion method, which extracts both local and interlocal spectral features, capturing thus also the correlations among nonadjacent bands. To our knowledge, this is the first reported deep spectral feature fusion method. Our model is a two-stream architecture, where an intergroup and a groupwise spectral classifier operate in parallel. The interlocal spectral correlation feature extraction is achieved elegantly, by reshaping the input spectral vectors to form the so-called nonadjacent spectral matrices. We introduce the concept of groupwise band convolution to enable the efficient extraction of discriminative local features with multiple kernels adopting the local spectral content. Another important contribution of this work is a novel dual-channel attention mechanism to identify the most informative spectral features. The model is trained in an end-to-end fashion with a joint loss. Experimental results on real datasets demonstrate excellent performance compared with the current state of the art.

Use of convolutional neural networks in total hip arthroplasty

Use of convolutional neural networks in total hip arthroplasty

Stacked Deep Dense Neural Network Model to Predict Alzheimer’s Dementia Using Audio Transcript Data

Alzheimer’s disease (AD) is caused by cortical degeneration leading to memory loss and dementia. A possible criterion for the early identification of Alzheimer’s dementia is to identify the difference between positive and negative linguistic and cognitive abilities of the patients. This study involves the use of Convolutional Neural Network (CNN), designed a hybrid model with CNN & Bidirectional Long-Short Term Memory (Bidirectional LSTM), and proposed a Stacked Deep Dense Neural Network (SDDNN) model for text classification and prediction of Alzheimer’s dementia. These models were trained end-to-end using DementiaBank clinical transcript dataset. The transcripts consisted of recorded interviews of Alzheimer’s patients with clinical experts. The models were investigated under two settings: Randomly initialized and Glove embedding. Further, hyperparameter optimization was accomplished using GridSearch, which yielded optimal parameters for the design of suitable learning models for most accurate predictions. Other parameters were computed and compared based on AUC, accuracy, specificity, precision, F1 score, and recall. To ensure performance generalization, the classification accuracy was tested using 10-fold cross-validation approach. The performance and classification accuracy of the proposed model was significantly improved to 93.31% when applied with Glove embedding and hyperparameter tuning. This research work will considerably help the clinical experts in early detection and diagnosis of AD.

A Modular System Based on U-Net for Automatic Building Extraction from very high-resolution satellite images

Recently, convolutional neural networks have grown in popularity in a variety of fields, such as computer vision and audio and text processing. This importance is due to the performance of this type of neural network in the state of the art, and in a wide variety of disciplines. However, the use of convolutional neural networks has not been widely used for remote sensing applications until recently. In this paper, we propose a CNN-based system capable of efficiently extracting buildings from very high-resolution satellite images, by combining the performances of the two architectures; U-Net and VGG19, which is obtained by putting two blocks in parallel based mainly on U-Net: The first block is a standard U-Net, and the second is designed by replacing the contraction path of standard U-Net with the pre-trained weights of VGG19.

Based on Artificial Intelligence Neural Network CNN Method Analysis and Processing of Dynamic Optical Breast Lesion Images

Breast cancer is the most common cancer in women. At present, the methods of examining lesions are generally mammography, or B-scanning and other methods with certain radioactive sources, which may lead to aggravation of breast lesions in young women. The dynamic optical breast scanning method uses infrared light to avoid the harm of X-rays to the human body. According to current research, doctors can use Dynamic Optical Breast Image (DOBI) to determine whether a patient has breast cancer. Studies have shown that convolutional neural networks (CNN) have higher detection accuracy in determining whether a patient has breast cancer. In this paper, we use an artificial intelligence neural network approach to analyze and process dynamic optical breast lesion images: we model the clinical lesion breast images in 3D, and use the VoxelMorph algorithm to segment the 3D images into 2D images; The time, space, location, and pathological trend curves in the image are analyzed and processed. We compared classification, sensitivity, and specific characteristics with the original dynamic breast lesion image scoring analysis system. The experimental results show that the accuracy is improved. At the same time, the problem of the original system's signature in the ROI area and leaf curve is solved. The use of CNN improves the analysis and processing speed, shortens the processing time, and increases the accuracy of the diagnostic reference from 83% to 90%.

AUTOMATED SYSTEM OF LUNG DISEASE DIAGNOSIS ON X-RAY IMAGES

Currently, automated detection of pneumonia in images is a priority in information technology. A promising option for solving this problem is the use of convolutional neural networks. The purpose of the study is to develop an automated system for diagnosing pneumonia from X-rays. This paper presents research and development of software for the automated system for analysis and classification of X-ray images of the lungs in order to automatically determine the signs of the disease, in particular pneumonia, which is most relevant due to the COVID-19 pandemic. This work is an example of the creation of a decision support system that is designed to assist a doctor to make decisions, analyze X-ray images of the lungs, classify them, and also allows to store all the necessary information about patients in one repository. For the automated system software was developed using C# and the user interface development environment - WPF. During the implementation process there were used the MVVM architecture and ML.NET as a tool for the neural network setup. Software was created based on the developed mathematical model through the integrated training of neural networks. To fulfill the main goal of software development, convolutional neural networks were used. As a result of the experiments, the coefficient of correctly recognized images was obtained – 97 %, which is close to the coefficient of recognition by a doctor.

Deformable convolutions in multi-view stereo

The Multi-View Stereo (MVS) is a key process in the photogrammetry workflow. It is responsible for taking the camera's views and finding the maximum number of matches between the images yielding a dense point cloud of the observed scene. Since this process is based on the matching between images it greatly depends on the ability of features matching throughout different images. To improve the matching performance several researchers have proposed the use of Convolutional Neural Networks (CNNs) to solve the MVS problem. Despite the progress in the MVS problem with the usage of CNNs, the Video RAM (VRAM) consumption within these approaches is usually far greater than classical methods, that rely more on RAM, which is cheaper to expand than VRAM. This work then follows the progress made in CasMVSNet in the reduction of GPU memory usage, and further study the changes in the feature extraction process. The Average Group-wise Correlation is used in the cost volume generation, to reduce the number of channels in the cost volume, yielding a reduction in GPU memory usage without noticeable penalties in the result. The deformable convolutions are applied in the feature extraction network to augment the spatial sampling locations with learning offsets, without additional supervision, to further improve the network's ability to model transformations. The impact of these changes is measured using quantitative and qualitative tests using the DTU and the Tanks and Temples datasets. The modifications reduced the GPU memory usage by 32% and improved the completeness by 9% with a penalty of 6.6% in accuracy on the DTU dataset.

Learning English and Arabic question similarity with Siamese Neural Networks in community question answering services

In this paper, we tackle the task of similar question retrieval (QR) which is essential for Community Question Answering (cQA) and aims to retrieve historical questions that are semantically equivalent to the new queries. Over time, with the sharp increase of community archives and the accumulation of duplicated questions, the QR problem has become increasingly challenging due to the shortness of the community questions as well as the word mismatch problem as users can formulate the same query using different wording. Although many efforts have been devoted to address this problem, existing methods mostly relied on supervised models which significantly depend on massive training data sets and manual feature engineering. Such methods are chiefly constrained by their specificities that ignore the word order and do not capture enough syntactic and semantic information in questions. In this paper, we rely on Neural Networks (NNs) which use a deep analysis of words and questions to take into consideration the semantics as well as the structure of questions to predict the semantic text similarity. We propose a deep learning approach based on a Siamese architecture with Long Short-Term Memory (LSTM) networks, augmented with an attention mechanism to let the model give different words different attention while modeling questions. We also explore the use of Convolutional Neural Networks (CNN) nested within the Siamese architecture to retrieve relevant questions. Different similarity measures were tested to predict the semantic similarity between the pairs of questions. To evaluate the proposed approach, we conducted experiments on large-scale datasets in English and Arabic.

Refinement of polygonal grids using Convolutional Neural Networks with applications to polygonal Discontinuous Galerkin and Virtual Element methods

We propose new strategies to handle polygonal grids refinement based on Convolutional Neural Networks (CNNs). We show that CNNs can be successfully employed to identify correctly the “shape” of a polygonal element so as to design suitable refinement criteria to be possibly employed within adaptive refinement strategies. We propose two refinement strategies that exploit the use of CNNs to classify elements' shape, at a low computational cost. We test the proposed idea considering two families of finite element methods that support arbitrarily shaped polygonal elements, namely Polygonal Discontinuous Galerkin (PolyDG) methods and Virtual Element Methods (VEMs). We demonstrate that the proposed algorithms can greatly improve the performance of the discretization schemes both in terms of accuracy and quality of the underlying grids. Moreover, since the training phase is performed off-line and is independent of the differential model the overall computational costs are kept low.

Construction Technology and Quality Control of Power and Electrical Engineering Based on Convolutional Neural Network

In the context of the Internet era, more and more parties have begun to store, process, and analyze data, but the accompanying question is whether people are reasonable about the data under the impact of massive data, effective and efficient analysis, especially the problems faced in this project. This article aims to study the quality control problems faced by electric power and electrical engineering in the construction process through the use of convolutional neural networks. Under this idea, this article proposes a multilayer convolution method. The experimental results show that the use of the improved multilayer convolution method for the convolution method of the convolutional neural network can effectively improve the multiple analysis problems of small datasets in the construction of electric power and electrical engineering; in this way, the relevant data are analyzed; by controlling the quality of construction, the quality problem has been greatly improved. After comparison, it is concluded that the overall construction quality has increased by 35%.

Analysis of Algorithms for Searching Objects in Images Using Convolutional Neural Network

The problem of finding objects in images using modern computer vision algorithms has been considered. The description of the main types of algorithms and methods for finding objects based on the use of convolutional neural networks has been given. A comparative analysis and modeling of neural network algorithms to solve the problem of finding objects in images has been conducted. The results of testing neural network models with different architectures on data sets VOC2012 and COCO have been presented. The results of the study of the accuracy of recognition depending on different hyperparameters of learning have been analyzed. The change in the value of the time of determining the location of the object depending on the different architectures of the neural network has been investigated.

A Biometric AsymmetricCryptosystem Software Module Based on Convolutional Neural Networks

During the research, the analysis of the existing biometric cryptographic systems was carried out. Some methods that help to generate biometric features were considered and compared with a cryptographic key. For comparing compact vectors of biometric images and cryptographic keys, the following methods are analyzed: designing and training of bidirectional associative memory; designing and training of single-layer and multilayer neural networks. As a result of comparative analysis of algorithms for extracting primary biometric features and comparing the generated image to a private key within the proposed authentication system, it was found that deep convolutional networks and neural network bidirectional associative memory are the most effective approach to process the data. In the research, an approach based on the integration of a biometric system and a cryptographic module was proposed, which allows using of a generated secret cryptographic key based on a biometric sample as the output of a neural network. The RSA algorithm is chosen to generate a private cryptographic key by use of convolutional neural networks and Python libraries. The software authentication module is implemented based on the client-server architecture using various internal Python libraries. Such authentication system should be used in systems where the user data and his valuable information resources are stored or where the user can perform certain valuable operations for which a cryptographic key is required. Proposed software module based on convolutional neural networks will be a perfect tool for ensuring the confidentiality of information and for all information-communication systems, because protecting information system from unauthorized access is one of the most pressing problems. This approach as software module solves the problem of secure generating and storing the secret key and author propose combination of the convolutional neural network with bidirectional associative memory, which is used to recognize the biometric sample, generate the image, and match it with a cryptographic key. The use of this software approach allows today to reduce the probability of errors of the first and second kind in authentication system and absolute number of errors was minimized by an average of 1,5 times. The proportion of correctly recognized images by the comparating together convolutional networks and neural network bidirectional associative memory in the authentication software module increased to 96,97%, which is on average from 1,08 times up to 1,01 times The authors further plan a number of scientific and technical solutions to develop and implement effective methods, tools to meet the requirements, principles and approaches to cybersecurity and cryptosystems for provide integrity and confidentiality of information in experimental computer systems and networks.

Aircraft Engine Performance Monitoring and Diagnostics Based on Deep Convolutional Neural Networks

The rapid advancement of machine-learning techniques has played a significant role in the evolution of engine health management technology. In the last decade, deep-learning methods have received a great deal of attention in many application domains, including object recognition and computer vision. Recently, there has been a rapid rise in the use of convolutional neural networks for rotating machinery diagnostics inspired by their powerful feature learning and classification capability. However, the application in the field of gas turbine diagnostics is still limited. This paper presents a gas turbine fault detection and isolation method using modular convolutional neural networks preceded by a physics-driven performance-trend-monitoring system. The trend-monitoring system was employed to capture performance changes due to degradation, establish a new baseline when it is needed, and generatefault signatures. The fault detection and isolation system was trained to step-by-step detect and classify gas path faults to the component level using fault signatures obtained from the physics part. The performance of the method proposed was evaluated based on different fault scenarios for a three-shaft turbofan engine, under significant measurement noise to ensure model robustness. Two comparative assessments were also carried out: with a single convolutional-neural-network-architecture-based fault classification method and with a deep long short-term memory-assisted fault detection and isolation method. The results obtained revealed the performance of the proposed method to detect and isolate multiple gas path faults with over 96% accuracy. Moreover, sharing diagnostic tasks with modular architectures is seen as relevant to significantly enhance diagnostic accuracy.