Person image synthesis, e.g., pose transfer, is a challenging problem due to large variation and occlusion. Existing methods have difficulties predicting reasonable invisible regions fail decouple the shape style of clothing, which limits their applications on person editing. In this paper, we propose PISE, novel two-stage generative model for Image Synthesis Editing, able generate realistic images with desired poses, textures, or semantic layouts. For human first synthesize parsing map...

The active recognition of interesting targets has been a vital issue for synthetic aperture radar (SAR) systems. SAR methods are mainly grouped as follows: extracting image features from the target amplitude or matching testing samples with template ones according to scattering centers extracted complex data. For image-based methods, convolutional neural networks (CNNs) achieve nearly highest accuracy images acquired under standard operating conditions (SOCs), while center feature-based...

The water body detection in high-resolution synthetic aperture radar (SAR) images is a challenging task due to the changing interference caused by multiple imaging conditions and complex land backgrounds. Inspired excellent adaptability of deep neural networks (DNNs) structured modeling capabilities probabilistic graphical models, cascaded fully-convolutional network (CFCN) proposed improve performance SAR images. First, for resolution loss convolutions with large stride traditional...

In recent years, terahertz imaging systems and techniques have been developed gradually become a leading frontier field. With the advantages of low radiation clothing-penetrable, technology has widely used for detection concealed weapons or other contraband carried on personnel at airports secure locations. This paper aims to detect these items with deep learning method its well performance real-time speed. Based analysis characteristics images, an effective system is proposed in this paper....

As an important remote sensing means, synthetic aperture radar (SAR) has many superiorities to other sensors. How effectively detect and locate ships in SAR images is also a popular field. In previous ship detection research, most algorithms focus on detecting the horizontal bounding box of targets, which ignore rotation angle each ships. Thus, too much background noise results makes them difficult describe accurately. Inspired by powerful feature representation ability convolutional neural...

The trajectory data of vessel AIS (automatic identification system) has important theoretical and application value for information supporting decisions. However, large sizes lead to difficulties in storing, querying, processing. To solve the problems high compression ratio longtime consumption existing online algorithm, an SPM (scan-pick-move) algorithm added sliding window is proposed. In order better compress regarding efficiency, classical algorithm. reduce storage space, maximum offset...

Abstract Human pose transfer, as a misaligned image generation task, is very challenging. Existing methods cannot effectively utilize the input information, which often fail to preserve style and shape of hair clothes. In this paper, we propose an adaptive human transfer network with two hierarchical deformation levels. The first level generates semantic parsing aligned target pose, second final textured person in guidance. To avoid drawback vanilla convolution that treats all pixels valid...

Automatic target recognition has been one of the hottest research in synthetic aperture radar (SAR) data processing. Noticing that popular methods cannot utilize multiple features SAR complex data, a method fused scattering center feature and deep convolutional neural network (CNN) is proposed this letter. This contains three key parts, namely, extraction reconstruction block, CNN final fusion classification block. In process, are at level maps, which retain space information 2-D maps. What...

Solar-induced chlorophyll fluorescence (SIF) is a hopeful indicator, which along with remote sensing, used to measure the photosynthetic efficiency and gross primary production (GPP) of vegetation in regional terrestrial ecosystems. Studies have found significant linear correlation between SIF GPP variety However, this relationship has mainly been established using data derived from satellite sensing continuous ground-based observations, respectively, are difficult accurately match. To...

The two-phase flow in a microchannel consists of liquid-liquid and gas-liquid material components. automatic recognition patterns using deep learning approaches has been emerging. This study aimed to improve the accuracy images. different convolutional kernels GoogLeNet algorithm extracted image features with scales. In order strengthen important channel spatial features, this paper proposes combined five-layer Coord attention enhance new algorithm. optimized model was derived from datasets...

3D human body representation learning has received increasing attention in recent years. However, existing works cannot flexibly, controllably and accurately represent bodies, limited by coarse semantics unsatisfactory capability, particularly the absence of supervised data. In this paper, we propose a with fine-grained high reconstruction-accuracy an unsupervised setting. Specifically, establish correspondence between latent vectors geometric measures parts designing part-aware...

As a typical application of remote sensing technology, change detection can find the ground information changes by acquiring images same region at different times. The using synthetic aperture radar (SAR) with advantages all day and all-weather usually monitors significant surface change, like flood disasters earthquake deformation. However, when it comes to detecting subtle vehicle traces, traditional methods ignoring phase coherence between image pairs cannot intensify these faint in...

To address the issue of insufficient safety monitoring power maintenance vehicles during operations, this study proposes a vehicle scheme based on ultra wideband (UWB) and deep learning. The UWB localization algorithm employs Chaotic Particle Swarm Optimization (CSPO) to optimize Time Difference Arrival (TDOA)/Angle (AOA) locating in order overcome adverse effects non-visual distance multipath substations significantly improve positioning accuracy vehicles. solve problem large aspect ratio...

Change detection is a crucial remote sensing application because it can locate the interesting changed regions and provide corresponding time-series information with multitemporal images acquired in same region. Synthetic aperture radar (SAR), advantages of all-day all-weather conditions, achieve high-resolution imaging from long operation distances. Moreover, coherence characteristics SAR system cause attained complex-value to be more sensitive ground surface features. Traditional...

Monitoring drought stress is crucial for estimating productivity and assessing the health status of forest ecosystems under global climate change. Solar-induced chlorophyll fluorescence (SIF) mechanistically coupled to photosynthesis has advantages over greenness-based vegetation indices in detecting drought. In recent years, SIF commonly been used monitoring crop ecosystems. However, response tower-based remains unclear. this study, we investigated potential monitor drought, which was...

Ground-based synthetic aperture radar interferometry (GB-InSAR) is a valuable tool for deformation monitoring. The 2D interferograms obtained by GB-InSAR can be integrated with 3D terrain model to visually and accurately locate deformed areas. process has been preliminarily realized geometric mapping assisted terrestrial laser scanning (TLS). However, due the line-of-sight (LOS) monitoring, shadow layover often occur in topographically rugged areas, which makes it difficult distinguish...

Abstract Trajectory outlier detection is one of the most important branches data mining topics. Most existing algorithms only utilized location trajectory points and neglected some factors such as speed, acceleration, corner. In this paper, we propose a algorithm based on dynamic differential threshold, called TODDT. Considering influence target motion information in period time, TODDT improves traditional distance-based by combining threshold into to discover more meaningful outliers. The...