Formulated as a pixel-level labeling task, data-driven neural segmentation models for cloud and corresponding shadow detection have achieved promising accomplishment in remote sensing imagery processing. The limited capability of these methods to delineate the boundaries clouds shadows, however, is still referred central issue precise detection. In this paper, we focus on rough location fine-grained boundary refinement dataset Landsat8 OLI therefore propose Refined UNet achieve goal. To end,...

Abstract The prevalence of dual usage and the relatively low cessation rate among e-cigarette (EC) users suggest that ECs have not demonstrated significant effectiveness as a smoking tool. Furthermore, there has been substantial increase in EC recent years. Therefore, objective this study is to investigate association between use incidence respiratory symptoms chronic obstructive pulmonary disease (COPD). A total 10,326 participants aged 20 55 years, without any diseases or COPD, were...

In recent years smart sport equipments have achieved great success in professional and amateur sports, as well body sensory systems; now discovering interesting knowledge the surge of data from those embedded sensors used sports is necessary focus our research. this paper, we investigate golf swing classification method based on deep convolutional neural network (deep CNN) fed with multi-sensor signals. Our club integrates two orthogonally affixed strain gage sensors, 3-axis accelerometer...

Haze removal is still an essential prerequisite for image processing and computer vision tasks, joint inference refinement of transmission maps remain challenging in the physical scattering model-based haze methods. In this article, we propose end-to-end learnable dehazing network, which referred to as Guided-Pix2Pix, jointly estimate refine map further dehaze images by equation. Instead a two-stage model predicting postprocessing transmission, Guided-Pix2Pix concatenates trainable Pix2Pix...

Epidemiological evidence of the effect high-level air pollution and its interaction with meteorological factors on risk acute exacerbation chronic obstructive pulmonary disease (AECOPD) is limited. Daily data AECOPD cases, pollutants were collected from 2015 to 2018 in Shijiazhuang. A distributed lag non-linear model (DLNM) was used explore cumulative PM2.5 AECOPD. The between estimated by a generalized additive (GAM) stratification model. total 4766 patients enrolled. After controlling for...

LoRa has been widely used in the Internet of Things (IoT) recently for its advantages long distance and low power consumption. However, security problem restricts practical application scenarios. The existing Advanced Encryption Standard (AES) technology high computational complexity energy consumption terminals with limited hardware resources. Therefore, this letter proposes a novel physical layer encryption algorithm based on modulation characteristics, which is suitable low-cost...

Abstract. Shape is an important aspect of spatial attributes land use segments in remotely sensed imagery, but it still rarely used as a component classification or image-based analysis. This study aims to quantitatively characterize classes using shape metrics. The conducted case area located south China, covering twelve scenes SPOT-5 images. There were total ten metrics selected for the analysis, namely, Convexity (CONV), Solidity (SOLI), Elongation (ELONG), Roundness (ROUND), Rectangular...

The use of smart sports equipment and body sensory systems supervising daily training is gradually emerging in professional amateur sports; however, the problem processing large amounts data from sensors used sport discovering constructive knowledge a novel topic focus our research. In this article, we investigate golf swing classification methods based on varieties representative convolutional neural networks (deep networks) which are fed with embedded multi-sensors, to group multi-channel...

Inpainting refers to reconstruct the incomplete image or video via analysing their context, feature of tailing etc. Convolutional neural network with deep learning is proved be an effective method achieve inpainting. However, those algorithms existed now usually have vague and blurry results huge amount time train models. To address this issue, article based on construction Context Encoders, continue use strategy combining encoders generative adversarial networks (GANs), which we add global...

Human motion and gesture recognition receive much concern in sports field, such as physical education fitness for all. Although plenty of mature applications appear training using photography, video camera, or professional sensing devices, they are either expensive inconvenient to carry. MEMS devices would be a wise choice students ordinary body builders portable have many built-in sensors. In fact, hand gestures is discussed studies inertial sensors based on similarity matching. However,...

Clustering by fast search and find of density peaks (CFSFDP) was proposed to create clusters finding high-density peaks, quickly. CFSFDP mainly based on two rules: 1) a cluster center has high dense point 2) lies at large distance from other centers. The effecti veness highly depends upon the cutoff (Cd), which is used estimate each data point. However, there need provide predefined Cd. In this paper, we propose an adaptive way accurate Cd using characteristics Improved Sheather-Jones (ISJ)...

In this paper we propose a blockchain-based cross-domain authentication strategy. This strategy uses the cosmos network model to enable mobile devices reliably access external domain networks when moving across domains. Our test results prove feasibility of and have better performance than other schemes.

In recent years, smart sports equipment and body sensor systems have become popular in professional amateur sports. One of a few remaining problems real-time applications is the discovery knowledge from embedded sensors data. training, such helps accelerated motor learning. The authors start with exploring possibilities classification golf swing performance 1-D convolutional neural network (CNN) real-time. They thoroughly investigate multiple data classifiers based on CNNs fed multi-sensor...

Cloud and shadow detection is an essential prerequisite for further remote sensing processing, whereas edge-precise segmentation remains a challenging issue. In Refined UNet, we considered the aforementioned task proposed two-stage pipeline to achieve segmentation. The isolated regions in however, bring inferior visualization should be sufficiently eliminated. Moreover, end-to-end model also expected jointly predict refine results. this paper, propose UNet v2 joint prediction refinement of...

Cache-enabled small base stations (SBS) are capable of relieving the heavy burden backhaul link and reducing transmission latency. The hit probability depends on coverage caching placement probabilities. However, interference in small-cell networks may significantly degrade probability. In this paper, for MIMO consisting SBS users, where both them equipped with multiple antennas, a joint alignment (IA) probabilistic (JIA-ProbC) scheme is proposed. Using tools from stochastic geometry,...

Recently, recurrent neural networks have been extensively utilized to address a time-dependent system of linear equations (TDSLEs) with inequality systems. Nevertheless, these existing studies only limit the variable without considering constraints on its derivatives, which may be challenging accomplish given task in practical applications when additional are introduced. Beyond that, matrix pseudoinverse is performed, and non-negative slack variables introduced solution process, increases...

Remote sensing images are usually contaminated by opaque cloud and shadow regions when acquired, thus detections become one of the essential prerequisites for restoration objects interest underneath before further processing analysis. Such a detection issue can be considered an image segmentation task supported some cutting-edge machine learning techniques, but edge-precise performance is still challenging. We tried to achieve this presenting Refined UNets carry out in both pipeline...

LoRa has been widely used in the Internet of Things (IoT). However, since only supports single symbol transmission its bandwidth, bit rate is limited. This restricts future development LoRa, such as Smart Home, Vehicles, Satellite IoT, etc. Toward this end, based on energy convergence characteristic fractional domain, we propose a multiplexed (MuLoRa) which transmits multiple symbols simultaneously same bandwidth. By adding address bits, collision between MuLoRa avoided. Considering hardware...

Remote sensing images are usually contaminated by cloud and corresponding shadow regions, making detection one of the essential prerequisites for processing translation remote images. Edge-precise segmentation remains challenging due to inherent high-level semantic acquisition current neural fashions. We, therefore, introduce Refined UNet series partially achieve edge-precise detection, including two-stage UNet, v2 with a potentially efficient gray-scale guided Gaussian filter-based CRF, v3...

In recent years, LoRa has been extensively researched in the satellite Internet of Things (IoT). However, multiple access technology is still one bottlenecks IoT. To improve performance IoT, based on orthogonality symbols fractional domain, this paper proposes a low complexity Orthogonal Multiple Access (OLMA) algorithm for users occupying same frequency bandwidth. The introduces address code to divide bandwidth into parts, and OLMA with different codes occupy parts transmit information...