Classification of Alzheimer's disease from MRI Images using CNN based Pre-trained VGG-19 Model

Determining the size of the tumor is a significant obstacle in brain tumour preparation and objective assessment. Magnetic Resonance Imaging (MRI) is one of the non-invasive methods that has emanated without ionizing radiation as a front-line diagnostic method for brain tumour. Several approaches have been applied in modern years to segment MRI brain tumours automatically. These methods can be divided into two groups based on conventional learning, such as support vectormachine (SVM) and random forest, respectively hand-crafted features and classifier method. However, after deciding hand-crafted features, it uses manually separated features and is given to classifiers as input. These are the time consuming activity, and their output is heavily dependent upon the experience of the operator. This research proposes fully automated detection of brain tumor using Convolutional Neural Network (CNN) to avoid this problem. It also uses brain image of high grade gilomas from the BRATS 2015 database. The suggested research performs brain tumor segmentation using clustering of k-means and patient survival rates are increased with this proposed early diagnosis of brain tumour using CNN.

Hybrid Convolutional Neural Network with PSO Based Severe Dengue Prognosis Method in Human Genome Data

Dengue is one of the most significant diseases transmitted by arthropods in the world. Dengue phenotypes are focused on documented inaccuracies in the laboratory and clinical studies. In countries with a high incidence of this disease, early diagnosis of dengue is still a concern for public health. Deep learning has been developed as a highly versatile and accurate methodology for classification and regression, which requires small adjustment, interpretable results, and the prediction of risk for complex diseases. This work is motivated by the inclusion of the Particle Swarm Optimization (PSO) algorithm for the fine-tuning of the model's parameters in the convolutional neural network (CNN). The use of this PSO was used to forecast patients with extreme dengue, and to refine the input weight vector and CNN parameters to achieve anticipated precision, and to prevent premature convergence towards local optimum conditions.

A Review on Prostate Cancer Detection using Deep Learning Techniques

The second most diagnosed disease of men throughout the world is Prostate cancer (PCa). 28% of cancers in men result in the prostate, making PCa and its identification an essential focus in cancer research. Hence, developing effective diagnostic methods for PCa is very significant and has critical medical effect. These methods could improve the advantages of treatment and enhance the patients' survival chance. Imaging plays a significant role in the identification of PCa. Prostate segmentation and classification is a difficult process, and the difficulties fundamentally vary with one imaging methodology then onto the next. For segmentation and classification, deep learning algorithms, specifically convolutional networks, have quickly become an optional technique for medical image analysis. In this survey, various types of imaging modalities utilized for diagnosing PCa is reviewed and researches made on the detection of PCa is analyzed. Most of the researches are done in machine learning based and deep learning based techniques. Based on the results obtained from the analysis of these researches, deep learning based techniques plays a significant and promising part in detecting PCa. Most of the techniques are based on computer aided detection (CAD) systems, which follows preprocessing, segmentation, feature extraction, and classification processes, which yield efficient results in detecting PCa. As a conclusion from the analysis of some recent works, deep learning based techniques are adequate for the detection of PCa.

A New Modified Recurrent Extreme Learning with PSO Machine Based on Feature Fusion with CNN Deep Features for Breast Cancer Detection

Breast cancer is a prevalent cause of death, and is the only form of cancer that is common among women worldwide and mammograms-based computer-aided diagnosis (CAD) program that allows early detection, diagnosis and treatment of breast cancer. But the performance of the current CAD systems is still unsatisfactory. Early recognition of lumps will reduce overall breast cancer mortality. This study investigates a method of breast CAD, focused on feature fusion with deep features of the Convolutional Neural Network (CNN). First, present a scheme of mass detection based on CNN deep features and modified clustering of the Extreme Learning Machine (MRELM). It forecasts load through Recurrent Extreme Learning Machine (RELM) and utilizes Artificial Bee Colony (ABC) to optimize weights and biases. Second, a collection of features is constructed that relays deep features, morphological features, texture features, and density features. Third, MRELM classifier is developed to distinguish benign and malignant breast masses using the fused feature set. Extensive studies show the precision and efficacy of the proposed method of mass diagnosis and classification of breast cancer.

Performance Analysis of Deep Belief Neural Network for Brain Tumor Classification

The brain tumors are by far the most severe and violent disease, contributing to the highest degree of a very low life expectancy. Therefore, recovery preparation is a crucial step in improving patient quality of life. In general , different imaging techniques such as computed tomography ( CT), magnetic resonance imaging ( MRI) and ultrasound imaging have been used to examine the tumor in the brain, lung , liver, breast , prostate ... etc. MRI images are especially used in this research to diagnose tumor within the brain with classification results. The massive amount of data produced by the MRI scan, therefore, destroys the manual classification of tumor vs. non-tumor in a given period. However for a limited number of images, it is presented with some constraint that is precise quantitative measurements. Consequently, a trustworthy and automated classification scheme is important for preventing human death rates. The automatic classification of brain tumors is a very challenging task in broad spatial and structural heterogeneity of the surrounding brain tumor area. Automatic brain tumor identification is suggested in this research by the use of the classification with Deep Belief Network (DBN). Experimental results show that the DBN archive rate with low complexity seems to be 97 % accurate compared to all other state of the art methods.

A New Neural Network-Based Intrusion Detection System for Detecting Malicious Nodes in WSNs

The Wireless Sensor Networks (WSNs) are vulnerable to numerous security hazards that could affect the entire network performance, which could lead to catastrophic problems such as a denial of service attacks (DoS). The WSNs cannot protect these types of attacks by key management protocols, authentication protocols, and protected routing. A solution to this issue is the intrusion detection system (IDS). It evaluates the network with adequate data obtained and detects the sensor node(s) abnormal behavior. For this work, it is proposed to use the intrusion detection system (IDS), which recognizes automated attacks by WSNs. This IDS uses an improved LEACH protocol cluster-based architecture designed to reduce the energy consumption of the sensor nodes. In combination with the Multilayer Perceptron Neural Network, which includes the Feed Forward Neutral Network (FFNN) and the Backpropagation Neural Network (BPNN), IDS is based on fuzzy rule-set anomaly and abuse detection based learning methods based on the fugitive logic sensor to monitor hello, wormhole and SYBIL attacks.

Face Recognition Framework based on Convolution Neural Network with modified Long Short Term memory Method

For real-world applications, such as video monitoring, interaction between human machines and safety systems, face recognition is very critical. Deep learning approaches have demonstrated better results in terms of precision and processing speed in image recognition compared to conventional methods. In comparison to traditional methods. While facial detection problems with different commercial applications have been extensively studied for several decades, they still face problems with many specific scenarios, due to various problems such as severe facial occlusions, very low resolutions, intense lighting and exceptional changes in image or video compression artifacts, etc. The aim of this work is to robustly solve the issues listed above with a facial detection approach called Convolution Neural Network with Long short-term Model (CNN-mLSTM). This method first flattened the original frame, calculating the gradient image with Gaussian filter. The edge detection algorithm Canny-Kirsch Method will then be used to identify edge of the human face. The experimental findings suggest that the technique proposed exceeds the current modern methods of face detection.

Classification of Diabetic Retinopathy using Stacked Autoencoder-Based Deep Neural Network

Diagnosis of diabetic retinopathy (DR) via images of colour fundus requires experienced clinicians to determine the presence and importance of a large number of small characteristics. This work proposes and named Adapted Stacked Auto Encoder (ASAE-DNN) a novel deep learning framework for diabetic retinopathy (DR), three hidden layers have been used to extract features and classify them then use a Softmax classification. The models proposed are checked on Messidor's data set, including 800 training images and 150 test images. Exactness, accuracy, time, recall and calculation are assessed for the outcomes of the proposed models. The results of these studies show that the model ASAE-DNN was 97% accurate.

A Hybrid Genetic-Neuro Algorithm for Cloud Intrusion Detection System

The security for cloud network systems is essential and significant to secure the data source from intruders and attacks. Implementing an intrusion detection system (IDS) for securing from those intruders and attacks is the best option. Many IDS models are presently based on different techniques and algorithms like machine learning and deep learning. In this research, IDS for the cloud computing environment is proposed. Here in this model, the genetic algorithm (GA) and back propagation neural network (BPNN) is used for attack detection and classification. The Canadian Institute for Cyber-security CIC-IDS 2017 dataset is used for the evaluation of performance analysis. Initially, from the dataset, the data are preprocessed, and by using the genetic algorithm, the attack was detected. The detected attacks are classified using the BPNN classifier for identifying the types of attacks. The performance analysis was executed, and the results are obtained and compared with the existing machine learning-based classifiers like FC-ANN, NB-RF, KDBN, and FCM-SVM techniques. The proposed GA-BPNN model outperforms all these classifying techniques in every performance metric, like accuracy, precision, recall, and detection rate. Overall, from the performance analysis, the best classification accuracy is achieved for Web attack detection with 97.90%, and the best detection rate is achieved for Brute force attack detection with 97.89%.

A Survey on Security Risks in Internet of Things (IoT) Environment

This analysis reviews the management of vulnerabilities and security risks of Internet of Things (IoT). This paper provides an overview, which it reveals the recent Internet's growth and how it has transformed our lives in various, unforeseen dimensions and how it has given rise to IoT. The introduction part focuses on providing an analysis on literature by presenting a short IoT history, some technical information on security protocols, and IoT hardware problems. The section on survey is where similar literatures on specific concepts are reviewed by describing the vulnerabilities and threats of IoT systems, and then reviewed risk management mechanisms for both information technologies and information protection. After the review, the analysis and discussion segment addressed and evaluated the details contained in the literature review. In this paper, a new risk management strategy uniquely designed for each IoT system is proposed. Then proposed work is evaluated by discussing the advantages and concluded the analysis and the future work.