Remote sensing continues to be an invaluable tool in earthquake damage assessments and emergency response. This study evaluates the effectiveness of multilayer feedforward neural networks, radial basis Random Forests detecting caused by 2010 Port-au-Prince, Haiti 7.0 moment magnitude (Mw) event. Additionally, textural structural features including entropy, dissimilarity, Laplacian Gaussian, rectangular fit are investigated as key variables for high spatial resolution imagery classification....

Conventional speaker recognition systems perform poorly under noisy conditions. Inspired by auditory perception, computational scene analysis (CASA) typically segregates speech producing a binary time-frequency mask. We investigate CASA for robust identification. first introduce novel feature, gammatone frequency cepstral coefficient (GFCC), based on an periphery model, and show that this feature captures characteristics performs substantially better than conventional features To deal with...

We propose and study a nonoverlapping nonconforming domain decomposition method for the integral-equation-based solution of large, complex electromagnetic (EM) scattering problems. The continuity electric surface current across boundary between adjacent subdomains is enforced by skew-symmetric interior penalty formulation. A additive Schwarz preconditioner designed analyzed linear system equations resulting from Galerkin boundary-element discretization. show that preconditioned exhibits...

A photoinduced deuterodetriazenation of aryltriazenes with CDCl3 under catalyst-free conditions is reported. The reactions featured simple operation, ecofriendly conditions, readily available reagents, inexpensive D sources, precise site selectivity, and a wide range substrates. Since could be synthesized from arylamine, this protocol can used as general method for easily accurately incorporating deuterium into aromatic systems by using source.

A conventional automatic speech recognizer does not perform well in the presence of noise, while human listeners are able to segregate and recognize noisy conditions. We study a novel feature based on an auditory periphery model for robust recognition. Specifically, gammatone frequency cepstral coefficients derived by applying analysis filterbank responses. Our evaluations show that proposed performs considerably better than acoustic features. further demonstrate integrating with...

The performance of speaker recognition systems drop significantly under noisy conditions. To improve robustness, we have recently proposed novel auditory features and a robust system using front-end based on computational scene analysis. In this paper, further study the by exploring different feature dimensions incorporating dynamic features. addition, evaluate in identification task number We find that one performs substantially better than conventional feature. Furthermore, our achieves...

The spectral confusion between shadow and water (or other dark surfaces) often results in suboptimal urban classification performances, especially from high-resolution satellite imagery (e.g. IKONOS). A method was developed to incorporate spatial indices of image objects improve the shadow/water detection. number indices, such as size, shape neighbour objects, were characterized differentiate objects. This generated superior detection performance compared a traditional per-field Extraction...

Advancing the sustainable use and conservation of marine environments is urgent. Tons debris including macro- microplastics generated on land are entering oceans, resources decreasing, many species facing extinction. Though satellite remote sensing techniques commonly used for global environmental monitoring, it still difficult to detect small objects such as floating vast ocean surface, ecosystems deep in oceans where light does not reach unobservable. An autonomous monitoring system...

This paper describes our convolutional neural network (CNN) system for Semantic Textual Similarity (STS) task. We calculated semantic similarity score between two sentences by comparing their vectors. generated vector of every sentence max pooling dimension word There are mainly trick points in system. One is that we trained a CNN to transfer GloVe vectors more proper form STS task before pooling. Another fully-connected (FCNN) difference probability score. decided all hyper parameters...

This research evaluated the potential for using MODIS Normalized Difference Vegetation Index (NDVI) 16-day composite (MOD13Q) 250 m time-series data to develop an annual crop type mapping capability throughout 480,000 km 2 Great Lakes Basin (GLB). An ecoregion-stratified approach was developed a two-step processing that included initial differentiation of cropland versus non-cropland and subsequent identification individual types. Major types were mapped calendar years 2002 2007. National...

A human listener has the ability to follow a speaker's voice while others are speaking simultaneously; in particular, can organize time-frequency energy of same speaker across time into single stream. In this paper, we focus on sequential organization cochannel speech, or mixtures two voices. We extract minimally corrupted segments, usable speech using robust multipitch tracking algorithm. The extracted is shown capture characteristics and improves identification (SID) performance various...

Conventional speaker recognition systems perform poorly under noisy conditions. Recent research suggests that binary time-frequency (T-F) masks be a promising front-end for robust recognition. In this paper, we propose novel auditory features based on an periphery model, and show these capture significant characteristics. Additionally, estimate uncertainties of the T-F masks, calculate likelihood scores using uncertainty decoding. Our approach achieves substantial performance improvement in...

Linear precoding (LP) techniques for a multiuser multiple-input-multiple-output broadcast channel is investigated and analytical results of the achievable sum capacity error performance zero forcing (ZF) linear precoders presented. It shown that detection signal-to-noise ratio ZF-LP can be accurately approximated by shrinking Chi-square distribution. The symbol rate its diversity gain are given. Then, an improved precoder based on minimum mean-square (MMSE) criterion derived. Its analysed...

This research examined sub-pixel land-cover classification performance for tree canopy, impervious surface, and cropland in the Laurentian Great Lakes Basin (GLB) using both time-series MODIS (Moderate Resolution Imaging Spectro radiometer) NDVI (Normalized Difference Vegetation Index) surface reflectance data. Classification training strategies included an entire-region approach ecoregion-stratified approach, multi-layer perceptron neural network classifiers. Although large variations...

Orchard machinery autonomous navigation is helpful for improving the efficiency of fruit production and reducing labor costs. Path planning one core technologies orchard machinery. As normally planted in straight parallel rows, trees are natural landmarks that can provide suitable cues intelligent This paper presents a novel method to realize path based on computer vision technologies. We combine deep learning least-square (DL-LS) algorithm carry out new line extraction scenarios. First,...

Advances in interconnect technologies, such as the increase of number metal layers and 3-D stacking technique, have paved way for higher functionality superior performance while reducing size, power, cost today's integrated circuit package products. However, whether or not preserves signal integrity (SI) has become a crucial concern system designers. In this study, systematic full-wave numerical approach, based on nonconformal finite-element domain decomposition method (DDM), is proposed...

Ever-increasing complexity in high-speed electronic devices and systems presents significant computational challenges the numerical analysis terms of desired accuracy, efficiency, scalable parallelism. The objective this paper is to investigate high-resolution, high-performance full-wave field solvers for electromagnetic simulations product-level integrated circuits (ICs) electronics. emphasis placed on advancing parallel algorithms that are provably facilitating a design-through-analysis...

Limited to the low spatial resolution of hyperspectral imaging sensor, mixed pixels are inevitable in images. Therefore, obtain endmembers and corresponding fractions pixels, unmixing becomes a hot spot field remote sensing. Endmember spectral variability (ESV), which is common images, affects accuracy. This paper proposes method based on maximum margin criterion derivative weights (MDWSU) reduce effect ESV unmixing. Firstly, MDWSU model, an effective fast algorithm employed for establishing...

Rural revitalization, against the political background of traditional cultural renaissance, was put forward as a national strategy. The countryside is not only mode residence but also way life. In order to explore its revitalization path and realize vision common prosperity, practice Jingmei Village in Dongkeng Town, Dongguan City, used multifunctional landscape medium reform from multiple aspects such compiling planning blueprint, sorting out regional context, creating batch rural features,...