CONCATENATION TECHNIQUE IN CONVOLUTIONAL NEURAL NETWORKS FOR COVID-19 DETECTION BASED ON X-RAY IMAGES
Yakoop Razzaz Hamoud Qasim, Habeb Abdulkhaleq Mohammed Hassan & Abdulelah Abdulkhaleq Mohammed Hassan, Taiz University, Yemen
In this paper we present a Convolutional Neural Network consisting of NASNet and MobileNet in parallel (concatenation) to classify three classes COVID-19, normal and pneumonia, depending on a dataset of 1083 x-ray images divided into 361 images for each class. VGG16 and RESNet152-v2 models were also prepared and trained on the same dataset to compare performance of the proposed model with their performance. After training the networks and evaluating their performance, an overall accuracy of 96.91%for the proposed model, 92.59% for VGG16 model and 94.14% for RESNet152. We obtained accuracy, sensitivity, specificity and precision of 99.69%, 99.07%, 100% and 100% respectively for the proposed model related to the COVID-19 class. These results were better than the results of other models. The conclusion, neural networks are built from models in parallel are most effective when the data available for training are small and the features of different classes are similar.
Deep Learning, Concatenation Technique, Convolutional Neural Networks, COVID-19, Transfer Learning.
Full Paper
https://aircconline.com/csit/papers/vol10/csit101602.pdf
Volume Link :
http://airccse.org/csit/V10N16.html
ARTIST, STYLE AND YEAR CLASSIFICATION USING FACE RECOGNITION AND CLUSTERING WITH CONVOLUTIONAL NEURAL NETWORKS
Doruk Pancaroglu STM A.S., Ankara, Turkey
Artist, year and style classification of fine-art paintings are generally achieved using standard image classification methods, image segmentation, or more recently, convolutional neural networks (CNNs). This works aims to use newly developed face recognition methods such as FaceNet that use CNNs to cluster fine-art paintings using the extracted faces in the paintings, which are found abundantly. A dataset consisting of over 80,000 paintings from over 1000 artists is chosen, and three separate face recognition and clustering tasks are performed. The produced clusters are analyzed by the file names of the paintings and the clusters are named by their majority artist, year range, and style. The clusters are further analyzed and their performance metrics are calculated. The study shows promising results as the artist, year, and styles are clustered with an accuracy of 58.8, 63.7, and 81.3 percent, while the clusters have an average purity of 63.1, 72.4, and 85.9 percent.
Face Recognition, Clustering, Convolutional Neural Networks, Art Identification.
For More Details :
https://aircconline.com/csit/papers/vol10/csit101604.pdf
Volume Link :
http://airccse.org/csit/V10N16.html
A SMART INTERNET-OF-THINGS APPLICATION FOR SHOE RECOMMENDATIONS USING PRESSURE SENSOR AND RASPBERRY PI
VYutian Fan1, Yu Sun2 and Fangyan Zhang3, 1Milton Academy, USA, 2California State Polytechnic University, USA 3ASML, USA
Running is one of the most important and simple sports spanning various ages, which can train throughout body and muscle. For running, proper shoes not only improve runners’ performance but also protect them from injury to some extent. However, runners have difficulty in finding a pair of shoes which fit runners’ gait patterns and feet shape very well. The process of selection of shoes is not effective and necessarily accurate. In this paper, we propose a new tool which facilitates the process by employing electronic sensors to the insoles of shoes and collecting feet information for runner accurately. It is helpful for runners to find the best fit shoes.
Machine learning, Firebase, Mobile application, Model fitting.
For More Details :
https://aircconline.com/csit/papers/vol10/csit101214.pdf
Volume Link :
http://airccse.org/csit/V10N12.html
MULTI SCALE TEMPORAL GRAPH NETWORKS FOR SKELETON-BASED ACTION RECOGNITION
Tingwei Li1, Ruiwen Zhang2, Qing Li1, 1,2 Tsinghua University, China
Graph convolutional networks (GCNs) can effectively capture the features of related nodes and improve the performance of model. More attention is paid to employing GCN in Skeleton-Based action recognition. But existing methods based on GCNs have two problems. First, the consistency of temporal and spatial features is ignored for extracting features node by node and frame by frame. To obtain spatiotemporal features simultaneously, we design a generic representation of skeleton sequences for action recognition and propose a novel model called Temporal Graph Networks (TGN). Secondly, the adjacency matrix of graph describing the relation of joints are mostly depended on the physical connection between joints. To appropriate describe the relations between joints in skeleton graph, we propose a multi-scale graph strategy, adopting a full-scale graph, part-scale graph and core-scale graph to capture the local features of each joint and the contour features of important joints. Experiments were carried out on two large datasets and results show that TGN with our graph strategy outperforms state-of-the-art methods.
Skeleton-based action recognition, Graph convolutional network, Multi-scale graphs.
For More Details :
https://aircconline.com/csit/papers/vol10/csit101605.pdf
Volume Link :
http://airccse.org/csit/V10N16.html
INTRUSION DETECTION IN COMPUTER SYSTEMS BY USING ARTIFICIAL NEURAL NETWORKS WITH DEEP LEARNING APPROACHES
Sergio Hidalgo-Espinoza, Kevin Chamorro-Cupuerán and Oscar Chang-Tortolero University of Yachay Tech, Ecuador
Intrusion detection into computer networks has become one of the most important issues in cybersecurity. Attackers keep on researching and coding to discover new vulnerabilities to penetrate information security system. In consequence computer systems must be daily upgraded using up-to-date techniques to keep hackers at bay. This paper focuses on the design and implementation of an intrusion detection system based on Deep Learning architectures. As a first step, a shallow network is trained with labelled log-in [into a computer network] data taken from the Dataset CICIDS2017. The internal behaviour of this network is carefully tracked and tuned by using plotting and exploring codes until it reaches a functional peak in intrusion prediction accuracy. As a second step, an autoencoder, trained with big unlabelled data, is used as a middle processor which feeds compressed information and abstract representation to the original shallow network. It is proven that the resultant deep architecture has a better performance than any version of the shallow network alone. The resultant functional code scripts, written in MATLAB, represent a re-trainable system which has been proved using real data, producing good precision and fast response.
Artificial Neural Networks, Information Security, Deep Learning, intrusion detection & hacking attacks
For More Details :
: https://aircconline.com/csit/papers/vol10/csit101501.pdf
Volume Link :
http://airccse.org/csit/V10N15.html
T3-D OFFLINE SIGNATURE VERIFICATION WITH CONVOLUTIONAL NEURAL NETWORK
Na Tyrer1, Fan Yang1, Gary C. Barber1, Guangzhi Qu1, Bo Pang1 and Bingxu Wang1,2 1Oakland University, USA 2 Zhejiang Sci-Tech University, P.R.China
Signature verification is essential to prevent the forgery of documents in financial, commercial, and legal settings. There are many researchers have focused on this topic, however, utilizing the 3-D information presented by a signature using a 3D optical profilometer is a relatively new idea, and the convolutional neural network is a powerful tool for image recognition. The present research focused on using the 3 dimensions of offline signatures in combination with a convolutional neural network to verify signatures. It was found that the accuracy of the data for offline signature verification was over 90%, which shows promise for this method as a novel method in signature verification. .
Signature Verification, 3D Optical Profilometer, Convolutional Neural Network.
For More Details :
https://aircconline.com/csit/papers/vol10/csit101518.pdf
Volume Link :
http://airccse.org/csit/V10N15.html
NETWORK DEFENSE IN AN END-TO-END PARADIGM
William R. Simpson and Kevin E. Foltz The Institute for Defense Analyses (IDA), Alexandria, Virginia, USA
Network defense implies a comprehensive set of software tools to preclude malicious entities from conducting nefarious activities. For most enterprises at this time, that defense builds upon a clear concept of the fortress approach. Many of the requirements are based on inspection and reporting prior to delivery of the communication to the intended target. These inspections require decryption of packets when encrypted. This decryption implies that the defensive suite has access to the private keys of the servers that are the target of communication. This is in contrast to an end-to-end paradigm where known good entities can communicate directly with each other. In an end-to-end paradigm, maintaining confidentiality through unbroken end-toend encryption, the private key resides only with the holder-of-key in the communication and on a distributed computation of inspection and reporting. This paper examines a formulation that is pertinent to the Enterprise Level Security (ELS) framework. .
Appliance, end-to-end security model, ELS, network defenses, web server handlers.
For More Details :
https://aircconline.com/csit/papers/vol10/csit101414.pdf
Volume Link :
http://airccse.org/csit/V10N14.html
DEEP LEARNING ROLES BASED APPROACH TO LINK PREDICTION IN NETWORKS
Aman Gupta* and Yadul Raghav* Indian Institute of Technology (BHU), Varanasi, India 221–005
The problem of predicting links has gained much attention in recent years due to its vast application in various domains such as sociology, network analysis, information science, etc. Many methods have been proposed for link prediction such as RA, AA, CCLP, etc. These methods required hand-crafted structural features to calculate the similarity scores between a pair of nodes in a network. Some methods use local structural information while others use global information of a graph. These methods do not tell which properties are better than others. With an in-depth analysis of these methods, we understand that one way to overcome this problem is to consider network structure and node attribute information to capture the discriminative features for link prediction tasks. We proposed a deep learning Autoencoder based Link Prediction (ALP) architecture for the latent representation of a graph, unified with non-negative matrix factorization to automatically determine the underlying roles in a network, after that assigning a mixed-membership of these roles to each node in the network. The idea is to transfer these roles as a feature vector for the link prediction task in the network. Further, cosine similarity is applied after getting the required features to compute the pairwise similarity score between the nodes. We present the performance of the algorithm on the real-world datasets, where it gives the competitive result compared to other algorithms. .
Link Prediction, Deep Learning, Autoencoder, Latent Representation, Non-Negative Matrix Factorization.
For More Details :
https://aircconline.com/csit/papers/vol10/csit101416.pdf
Volume Link :
http://airccse.org/csit/V10N14.html
ENERGY AWARE ROUTING WITH COMPUTATIONAL OFFLOADING FOR WIRELESS SENSOR NETWORKS
Adam Barker and Martin Swany , Indiana University, Bloomington, Indiana, USA
Wireless sensor networks (WSN) are characterized by a network of small, battery powered devices, operating remotely with no pre-existing infrastructure. The unique structure of WSN allow for novel approaches to data reduction and energy preservation. This paper presents a modification to the existing Q-routing protocol by providing an alternate action of performing sensor data reduction in place thereby reducing energy consumption, bandwidth usage, and message transmission time. The algorithm is further modified to include an energy factor which increases the cost of forwarding as energy reserves deplete. This encourages the network to conserve energy in favor of network preservation when energy reserves are low. Our experimental results show that this approach can, in periods of high network traffic, simultaneously reduce bandwidth, conserve energy, and maintain low message transition times. .
Ad Hoc Network Routing, Q-routing, Wireless Sensor Network, Computational Offloading, Energy Aware.
For More Details :
https://aircconline.com/csit/papers/vol10/csit101418.pdf
Volume Link :
http://airccse.org/csit/V10N14.html
DATA PREDICTION OF DEFLECTION BASIN EVOLUTION OF ASPHALT PAVEMENT STRUCTURE BASED ON MULTI-LEVEL NEURAL NETWORK
Shaosheng Xu1, Jinde Cao2 and Xiangnan Liu2, 1,2Southeast University, China
Aiming at reducing the high cost of test data collection of deflection basins in the structural design of asphalt pavement and shortening the long test time of new structures, this paper innovatively designs a structure coding network based on traditional neural networks to map the pavement structure to an abstract space. Therefore, the generalization ability of the neural network structure is improved, and a new multi-level neural network model is formed to predict the evolution data of the deflection basin of the untested structure. By testing the experimental data of RIOHTRACK, the network structure predicts the deflection basin data of untested pavement structure, of which the average prediction error is less than 5%. .
multi-level neural network, Encoding converter, structural of asphalt pavement, deflection basins, RIOHTRACK.
For More Details :
https://aircconline.com/csit/papers/vol10/csit101304.pdf
Volume Link :
http://airccse.org/csit/V10N13.html
STABILITY ANALYSIS OF QUATERNIONVALUED NEURAL NETWORKS WITH LEAKAGE DELAY AND ADDITIVE TIME-VARYING DELAYS
Qun Huang and Jinde Cao School of Mathematics, Southeast University, Nanjing, China
In this paper, the stability analysis of quaternion-valued neural networks (QVNNs) with both leakage delay and additive time-varying delays is proposed. By employing the LyapunovKrasovskii functional method and fully considering the relationship between time-varying delays and upper bounds of delays, some sufficient criteria are derived based on reciprocally convex method and several inequality techniques. The stability criteria are established in two forms: quaternion-valued linear matrix inequalities (QVLMIs) and complex-valued linear matrix inequalities (CVLMIs),in which CVLMIs can be directly resolved by the Yalmip toolbox in MATLAB. Finally, an illustrative example is presented to demonstrate the validity of the theoretical results. .
Quaternion-valued Neural Networks, Stability Analysis, Lyapunov-KrasovskiiFunctional, Leakage Delay, Additive Time-varying Delays
For More Details :
https://aircconline.com/csit/papers/vol10/csit101305.pdf
Volume Link :
http://airccse.org/csit/V10N13.html
QUALITY OF SERVICE-AWARE SECURITY FRAMEWORK FOR MOBILE AD HOC NETWORKS USING OPTIMIZED LINK STATE ROUTING PROTOCOL
Thulani Phakathi, Francis Lugayizi and Michael Esiefarienrhe North-West University, Mafikeng, South Africa
All networks must provide an acceptable and desirable level of Quality of Service (QoS) to ensure that applications are well supported. This becomes a challenge when it comes to Mobile ad-hoc networks (MANETs). This paper presents a security framework that is QoS-aware in MANETs using a network protocol called Optimized Link State Routing Protocol (OLSR). Security & QoS targets may not necessarily be similar but this framework seeks to bridge the gap for the provision of an optimal functioning MANET. This paper presents the various security challenges, attacks, and goals in MANETs and the existing architectures or mechanisms used to combat security attacks. Additionally, this framework includes a security keying system to ascertain QoS. The keying system is linked to the basic configuration of the protocol OLSR through its Multi-point Relays (MPRs) functionality. The proposed framework is one that optimizes the use of network resources and time. .
Routing protocols, MANETs, Trust framework, Video streaming, QoS.
For More Details :
https://aircconline.com/csit/papers/vol10/csit101109.pdf
Volume Link :
http://airccse.org/csit/V10N11.html
AUTOMATED CLASSIFICATION OF BANANA LEAF DISEASES USING AN OPTIMIZED CAPSULE NETWORK MODEL
Bolanle F. Oladejo and Oladejo Olajide Ademola University of Ibadan, Ibadan, Nigeria
Plant disease detection and classification have undergone successful researches using Convolutional Neural Network (CNN); however, due to the intrinsic inability of max pooling layer in CNN, it fails to capture the pose, view and orientation of images. It also requires large training data and fails to learn the spatial relationship of the features in an object. Thus, Capsule Network (CapsNet) is a novel deep learn- ing model proposed to overcome the shortcomings of CNN. We developed an optimized Capsule Network model for classification problem using banana leaf diseases as a case study. The two dataset classes in- clude Bacterial Wilt and Black Sigatoka, with healthy leaves. The developed model adequately classified the banana bacterial wilt, black sigatoka and healthy leaves with a test accuracy of 95%. Its outper- formed three variants of CNN architectures implemented (a trained CNN model from scratch, LeNet5 and ResNet50) with respect to rotation invariance. .
Capsule Network, CNN, Activation function, Deep Learning, Precision Agriculture..
For More Details :
https://aircconline.com/csit/papers/vol10/csit100910.pdf
Volume Link :
http://airccse.org/csit/V10N09.html
WIRELESS SENSOR NETWORKS SIMULATORS AND TESTBEDS
Souhila Silmi1,2, Zouina Doukha1, Rebiha Kemcha2,3 and Samira Moussaoui1 1USTHB University, Algeria 2Higher Normal School Elbachir Elibrahimi-Kouba, Algeria 3University of Boumerdes, Algeria
Wireless sensor networks (WSNs) have emerged as one of the most promising technologies for the current era. Researchers have studied them for several years ago, but more work still needed to be made since open opportunities to integrate new technologies are added to this field. One challenging task is WSN deployment. Yet, this is done by real deployment with testbeds platforms or by simulation tools when real deployment could be costly and timeconsuming. In this paper, we review the implementation and evaluation process in WSNs. We then describe relevant testbeds and simulation tools, and their features. Lastly, we conduct an experimentation study using these testbeds and simulations to highlight their pro and cons. As a use case, we implement a localization protocol. This work gives clarity to future-work for better implementation in order to improve reliability, accuracy and time consumed. .
Wireless Sensor Network, Testbeds, Simulation Tools, Localization Protocol.
For More Details :
https://aircconline.com/csit/papers/vol10/csit100912.pdf
Volume Link :
http://airccse.org/csit/V10N09.html
TIBETAN AND CHINESE TEXT IMAGE CLASSIFICATION BASED ON CONVOLUTIONAL NEURAL NETWORK
Jincheng Li1, Penghai Zhao1, Yusheng Hao2, Qiang Lin2, Weilan Wang1* 1Northwest Minzu University,P. R. China 2Northwest Minzu University, P. R. China
The first stage of Tibetan-Chinese bilingual scene text detection and recognition is the detection of Tibetan- Chinese bilingual scene text. The detection results are mainly divided into three categories: successfully detected regions of Tibetan text and Chinese text, non-words regions with failed predictions. If the detected text image results are accurately classified, then the nontext images should be filtered in the recognition phase, meanwhile the Tibetan and Chinese text images can be identified by using different classifiers, such procedure can reduce the complexity of classification and recognition of two different characters by one recognition model. An accurate classification of Tibetan and Chinese text images is mattered. Therefore, this paper conducts a research on the classification of Tibetan, Chinese and non-text images by using convolutional neural networks. We perform a series of exploration about the classification accuracy of Tibetan, Chinese text images and non-text images with convolutional neural networks in different depths, and compare the accuracy with the classification results based on the transfer learning then analyze it. The results show that for the classification of Tibetan, Chinese and non-text images in the scene, using 7-layer convolutional neural network has reached saturation, and increasing the network depth does not improve the results, which provides reference values for Tibetan-Chinese text image classification. .
Convolutional Neural Network, Tibetan-Chinese scene text image, image classification, transfer learning.
For More Details :
https://aircconline.com/csit/papers/vol10/csit100807.pdf
Volume Link :
http://airccse.org/csit/V10N08.html
BACK-PROPAGATION NEURAL NETWORKBASED METHOD FOR PREDICTING THE INTERVAL NATURAL FREQUENCIES OF STRUCTURES WITH UNCERTAIN-BUTBOUNDED PARAMETERS
Pengbo Wang1*, Wenting Jiang2 and Qinghe Shi3 1Beihang University, China 2Chinese Academy of Sciences, China 3Jiangsu University of Technology, China
Uncertain-but-bounded parameters have a significant impact on the natural frequencies of structures, and it is necessary to study their inherent relationship. However, their relationship is generally nonlinear and thus very complicated. Taking advantage of the strong non-linear mapping ability and high computational efficiency of BP neural networks, namely the error back-propagation neural networks, a BP neural network-based method is proposed to predict the interval natural frequencies of structures with uncertain-but-bounded parameters. To demonstrate the proposed method’s feasibility, a numerical example is tested. The lower and upper frequency bounds obtained using the proposed approach are compared with those obtained using the interval-based perturbation method, which is a commonly used method for problems with uncertainties. A Monte Carlo simulation is also conducted because it is always referred to as a reference solution for problems related to uncertainties. It can be observed that as the varying ranges of uncertain parameters become larger, the accuracy of the perturbation method deteriorates remarkably, but the proposed method still maintains a high level of accuracy. This study not only puts forward a novel approach for predicting the interval natural frequencies but also exhibits the broad application prospect of BP neural networks for solving problems with uncertainties. .
Back-propagation neural network, Natural frequency, Interval parameter, Perturbation method, Monte Carlo simulation.
For More Details :
https://aircconline.com/csit/papers/vol10/csit100702.pdf
Volume Link :
http://airccse.org/csit/V10N07.html
FCNNMD: A NOVEL FUSION METHOD BASED ON CONVOLUTIONAL NEURAL NETWORK FOR MALWARE DETECTION
Jing Zhang1 and Yu Wen2, 1University of Chinese Academy of Sciences, China, 2Chinese Academy of Sciences, China
Malicious software are rampant and do great harm. The present mainstream malware detection technology has many disadvantages, such as high labour cost, large system overhead, and inability to detect new malware. We propose a novel fusion method based on convolutional neural network for malware detection (FCNNMD). For the sample imbalance problem faced by the convolutional neural network malware detection method, the non-malicious sample is added by means of generating anti-network generation, etc., to achieve the same number as the malicious sample. For the problem of low accuracy of single model detection, high false positive rate and false negative rate, a malware detection model is constructed by means of model fusion. The model combines four classical convolutional neural network structures. Experiments show that this method can effectively improve the accuracy and robustness of the model. Our method does not need actual running software and has high detection efficiency. .
Malware Detection, Grayscale Image, Convolutional Neural Networks, Model integration.
For More Details :
https://aircconline.com/csit/papers/vol10/csit100601.pdf
Volume Link :
http://airccse.org/csit/V10N06.html
ONTOLOGY-BASED MODEL FOR SECURITY ASSESSMENT: PREDICTING CYBERATTACKS THROUGH THREAT ACTIVITY ANALYSIS
Pavel Yermalovich and Mohamed Mejri Faculté des Sciences et de Génie, Université Laval, Québec City, Canada
The prediction of attacks is essential for the prevention of potential risk. Therefore, risk forecasting contributes a lot to the optimization of the information security budget. This article focuses on the ontology and stages of a cyberattack. It introduces the main representatives of the attacking side and describes their motivation. .
Cyberattack, cyberattack prediction, ontology, cyberattack ontology, information security, cybersecurity,IT security, data security, threat activity.
For More Details :
https://aircconline.com/csit/papers/vol10/csit100305.pdf
Volume Link :
http://airccse.org/csit/V10N03.html
EVENT-BASED REAL-TIME HAND GESTURE RECOGNITION USING SPIKING NEURAL NETWORK
Van Khoa LE and Sylvain Bougnoux IMRA Europe S.A.S., 220 Rue Albert Caquot, 06904 Sophia-Antipolis
Deep learning represents the state of the art in many machine learning and computer vision problem. The core of this technology is the analog neural network (ANN) composed of multiple convolution and pooling layers. Unfortunately, such system demands massive computational power thus consuming a lot of energy and therefore causing negative effect to the environment. On one hand, human brain is known to be much more energy efficient, so the spiking neural network (SNN) was created to replicate the brain activity in order to improve the energy efficiency of current deep learning model. On the other hand, the event-based domain based on neuromorphic sensor like event camera made huge progress since last few years and become more and more popular. The data signal flow in spiking neural network is a perfect fit for the output of event camera. Therefore, in this article we built a system based on the combination of event camera and SNN for the real-time hand gesture recognition. We also give an analysis to prove the energy efficiency of this technology compared to the ANN counterpart. .
Event camera, Spiking Neural Network, Neuromorphic engineering.
For More Details :
https://aircconline.com/csit/papers/vol10/csit100105.pdf
Volume Link :
http://airccse.org/csit/V10N01.html
RGBD BASED GENERATIVE ADVERSARIAL NETWORK FOR 3D SEMANTIC SCENE COMPLETION
Jiahao Wang, Ling Pei*, Danping Zou, Yifan Zhu, Tao Li and Ruochen Wang Shanghai Key Laboratory of Location-based Navigation and Services SJTU-ParisTech Elite Institute of Technology Shanghai Jiao Tong University, Shanghai, China
3D scene understanding is of importance since it is a reflection about the real-world scenario. The goal of our work is to complete the 3d semantic scene from an RGB-D image. The state-ofthe-art methods have poor accuracy in the face of complex scenes. In addition, other existing 3D reconstruction methods use depth as the sole input, which causes performance bottlenecks. We introduce a two-stream approach that uses RGB and depth as input channels to a novel GAN architecture to solve this problem. Our method demonstrates excellent performance on both synthetic SUNCG and real NYU dataset. Compared with the latest method SSCNet, we achieve 4.3% gains in Scene Completion (SC) and 2.5% gains in Semantic Scene Completion (SSC) on NYU dataset. .
Scene Completion, Semantic Segmentation, Generation Adversarial Network, RGB-D.
For More Details :
https://aircconline.com/csit/papers/vol10/csit100111.pdf
Volume Link :
http://airccse.org/csit/V10N01.html
Journals by Area
Conference Publicity
To list your conference in this page, please contact us
