Multimodal image registration is the fundamental technique for scene analysis with series remote sensing images of different spectrum region. Due to highly nonlinear radiometric relationship, it quite challenging find common features between modal types. This paper resorts deep neural network, and tries learn descriptors multimodal patch matching, which key issue registration. A Siamese fully convolutional network set up trained a novel loss function, adopts strategy maximizing feature...

To evade malicious content detection, malware authors use packers, binary tools that instigate code obfuscation. By using executable modern can completely bypass personal firewalls and antivirus (AV) scanners.Reverse engineering (RE) has become an important approach to analyzing a program's logic flow internal data structures, such as system call functions. Security researchers AV products must be able unpack inspect the payloads hidden within packed programs RE tools.

Abstract In this paper, we propose a topological plasmonic 1D nanocavity, which comprises dielectric nanoparticle chain placed on metallic substrate. The structure is designed based the characteristics of photonic crystals. A edge state exists at interface between two distinct structures. comparison to non-topological defect mode, mode exhibits better properties including good confinement and robustness against perturbations in scales length-width ratio nanoparticles.

Named Data Networking (NDN) is a data-centric architecture designed for the future Internet. Existing works show that NDN brings significant performance improvement typical content-centric applications, and can also fit mobile environment well. However, directly applying to Vehicular Ad hoc NETworks (VANETs) confronted with great challenges due high mobility of vehicles. Most applications in VANETs are relied on data dissemination mechanisms. Therefore, we aim improve packet forwarding...

Change detection for synthetic aperture radar (SAR) images is a key process in many applications exploiting remote-sensing images. It challenging task due to the presence of speckle noise SAR imaging. This article investigates problem change multitemporal Our motivation avoid using only one detector measure level different features which usually considered by classical methods. In this article, we propose an unsupervised approach based on frequency difference wavelet domain and modified...

This letter presents an effective approach for synthetic aperture radar (SAR) image registration. Corresponding edge features are extracted by exploring the local spatial relationship around corresponding speeded-up robust feature points. Only pairs that matched in aspect of distance and orientation to center points taken as features. For edge-point set matching, we introduce coherent point drift (CPD) registration method. CPD is able maintain global topological structure edges outliers. We...

Due to the extremely complex composition of remote sensing scenes, REmote Sensing Image Scene Classification (RE-SISC) is still a challenging task.To further improve classification accuracy, this article introduces deep-learning detector into RE-SISC and proposes classify images according detected class-specific signature objects.Inspired by procedure human vision system, we design framework that utilizes objects scene classes guide classification.When performing image classification,...

Saliency, the distinctive parts of an image, has shown good potential for many applications (e.g., image interpretation and target detection). In this letter, a novel saliency detection method based on background context-aware is proposed synthetic aperture radar (SAR) images. According to statistical analysis SAR characteristics, several reference patches (RBPs) are selected. Then, dissimilarities between current patch RBPs used calculate local saliency, which further enhanced final under...

Non-negative matrix factorization (NMF) ignores both the local geometric structure of and discriminative information contained in a data set. A manifold geometry-based NMF dimension reduction method called discriminant (LDNMF) is proposed this paper. LDNMF preserves not only non-negativity but also data. The can be characterized by within-class graph between-class graph. An efficient multiplicative updating procedure produced, its global convergence guaranteed theoretically. Experimental...

A novel manifold learning feature extraction approach—preserving neighborhood discriminant embedding (PNDE) of hyperspectral image is proposed in this paper. The local geometrical and structure the data can be accurately characterized by within-class neighboring graph between-class graph. Unlike learning, such as LLE, Isomap LE, which cannot deal with new test samples images larger than 70×70, method here process full scene images. Experiments results on datasets real-word show that...

The scale parameter (SP), to control the sizes of objects, is great significance in multiscale segmentation which a prerequisite and foundational step for object-based change detection (OBCD). However, appropriate SP not readily apparent majority existing OBCD algorithms obtain SPs by empirical or subjective trial-and-error ways that may lead dissatisfactory accuracy be time-consuming. To address this issue, an automatic approach optimal selection proposed letter. First, changed fuzziness...

In this letter, an adaptively weighted multi-feature-based method for unsupervised object-based change detection in high-resolution remote sensing images is proposed. First, a sample selection strategy using fuzzy c-means designed to obtain high precision pseudo-samples way. Second, the multiple candidate features are categorized into spectral, geometric and textural groups two kinds of weights involved taking account different contributions. The within-group each feature can be calculated...

Customers always complain that anti-virus softwares bog down their computers by consuming much of PC memories and resources. With the popularity variety zero- day threats over Internet, security companies have to keep on inserting new virus signatures into databases. However, is increasing size signature file sole reason drag a crawl during scan? This paper outlines other three reasons for slowing software-protected computers, which actually are not directly related file. First, rising time...

SAR images despeckling is one of the most important research subjects in processing. Following discussion on advantage and disadvantage curvelet wavelet transforms based despeckling, an algorithm which combined was proposed for high resolution images. Firstly, were used to denoise original image, then related residual image between their results acquired, denoised through transform. Finally added image. Experimental TerraSAR-X indicated that, compared with Enhanced Lee filter, Frost...

Feature extraction acts a crucial role in application and research of hyperspectral data minimum noise fraction (MNF) is one the most common methods feature extraction. Estimation covariance matrix an inevitable step MNF, but it would bring error which lead imprecise while using first some components MNF transform to represent original data. To solve problem above, method based on singular value decomposition was proposed. acted data, then, transformed were decomposed by decomposition,...

A novel fast SAR image change detection method is presented in this paper. Based on a Bayesian approach, the prior information that speckles follow Nakagami distribution incorporated into difference (DI) generation process. The new DI performs much better than familiar log ratio (LR) as well cumulant based Kullback-Leibler divergence (CKLD) DI. statistical region merging (SRM) approach first introduced to context. clustering procedure with variance inference variable exhibited tailor...

Coherent point drift (CPD) method is a powerful registration tool under the framework of Gaussian mixture model (GMM). However, global spatial structure sets considered only without other forms additional attribute information. The equivalent simplification mixing parameters and manual setting weight parameter in GMM make CPD less robust to outlier have flexibility. An adaptive proposed automatically determine by embedding local information features into construction GMM. In addition,...

Abstract Background The accurate characterization of protein functions is critical to understanding life at the molecular level and has a huge impact on biomedicine pharmaceuticals. Computationally predicting function been studied in past decades. Plagued by noise errors protein–protein interaction (PPI) networks, researchers have undertaken focus fusion multi-omics data recent years. A model that appropriately integrates network topologies with biological preserves their intrinsic...