An Intrusion detection system is an essential security tool for protecting services and infrastructures of wireless sensor networks from unseen unpredictable attacks. Few works machine learning have been proposed intrusion in that achieved reasonable results. However, these still need to be more accurate efficient against imbalanced data problems network traffic. In this paper, we a new model detect attacks based on genetic algorithm extreme gradient boosting (XGBoot) classifier, called...

The firmness of meningiomas is a critical factor that impacts the surgical approach recommended for patients. conventional approaches couple image processing techniques with radiologists’ visual assessments magnetic resonance imaging (MRI) proved to be time-consuming and subjective physician’s judgment. Recently, machine learning-based methods have emerged classify MRI instances into firm or soft categories. Typically, such solutions rely on hand-crafted attributes and/or feature engineering...

Hyperspectral image classification has been increasingly used in the field of remote sensing. In this study, a new clustering framework for large-scale hyperspectral (HSI) is proposed. The proposed four-step scheme explores how to effectively use global spectral information and local spatial structure data HSI classification. Initially, multidimensional Watershed pre-segmentation. Region-based hierarchical segmentation based on construction Binary partition trees (BPT). Each segmented region...

AbstractWe propose novel generic performance metric for hyperspectral unmixing techniques. This relative compares two abundance matrices. The first one represents the result. second matrix can be either another result or ground truth of scene. starts by computing coincidence matrices corresponding to matrices, then comparison is carried out statistics number pairs data points that have high abundances with respect same endmember approach, but large differences technique, in both. main...

A physical verification flow of the layout silicon photonic circuits is suggested. Simple empirical models are developed to estimate bend power loss and coupled in integrated fabricated using SOI standard wafers. These utilized circuit verify reliable fabrication any electronic design automation tool. The accurate compared with electromagnetic solvers. closed form circumvent need utilize EM solver for process. Hence, it dramatically reduces time

We introduce a new spectral mixture analysis approach. The proposed Mixture Analysis based on Spectral Summarization (MASS) uses all the wavelengths of hyperspectral image and assumes convex geometry model in order to estimate endmembers their corresponding abundances. MASS unmixing technique is information provided by summarization image. performed through fuzzy partitioning scene.

Hyperspectral imagery is a main tool of remote sensing applications. As signal transmitted towards given scene, reflected and scattered again by interacting with the various components atmosphere surface, reflectance spectra analysis allows recognition and/or quantification materials. image three-dimensional data cube that containing values radiation has been collected over an area in wide range wavelengths. This hyperspectral serves to identify scene composition, includes applications such...

We propose a novel image database categorization approach using robust unsupervised learning of finite generalized dirichlet mixture models with feature discrimination. The proposed algorithm is based on optimizing an objective function that associates two types memberships each data sample. first one the posterior probability and indicates how well sample fits estimated distribution. second membership represents degree typicality used to identify discard noise points outliers. In addition,...

A simple analytical model is developed to estimate the power loss and time delay in photonic integrated circuits fabricated using SOI standard wafers. This can be utilized physical verification of circuit layout verify its feasibility for fabrication certain foundry specifications. allows providing new design rules process any electronic automation (EDA) tool. The accurate compared with finite element based full wave electromagnetic EM solver. closed form circumvents need utilize solver...

Recently, deep learning has been coupled with notice- able advances in natural language processing (NLP) related research. In this work, we propose a general framework to detect verbal offense social networks comments. We introduce partitional CNN-LSTM architecture order automatically recognize ver- bal patterns network Specifically, use CNN along LSTM model map the comments into two predefined classes. particular, rather than considering whole document/comments as input performed using...

In this paper, we propose novel generalized performance measures for hyperspectral unmixing techniques. Theses relative compare two abundances matrices. The first one represents the result. second matrix can be either another result or ground truth of scene. These start by computing coincidence matrices corresponding to abundance Then, comparison is carried out statistics number pairs data points that have high with respect same endmember approach, but large difference technique, in both....

We present accurate models compared to electromagnetic (EM) solvers for the coupling coefficient of two parallel waveguides and bend loss. These can be used with any electronic design automation (EDA) tool as equations depend only on waveguide dimensions which extracted easily from layout. Hence these will save a large amount time.

Recently, deep learning has been coupled with notice-able advances in Natural Language Processing related research.In this work, we propose a general framework to detect verbal offense social networks comments.We introduce partitional CNN-LSTM architecture order automatically recognize ver-bal patterns network comments.Specifically, use CNN along LSTM model map the comments into two predefined classes.In particular, rather than considering whole document/comments as input performed using...