Low-rank tensor completion methods have been advanced recently for modeling sparsely observed data with a multimode structure. However, low-rank priors may fail to interpret the model factors of general objects. The most common method address this drawback is use regularizations together priors. due complex nature and diverse characteristics real-world multiway data, single or few remains far from efficient, there are limited systematic experimental reports on advantages these completion. To...

Different from the ground image with uniform haze, haze in remote sensing (RS) has characteristics of irregular shape and uneven concentration hazy weather. It brings a great challenge to application RS data advanced processing tasks. A novel dehazing network for non-uniform image, named as KFA-Net, is proposed solve aforementioned issues. The designed asymmetric size feature cascade (ASFC), k-means pixel attention (KPA) FFT channel (FCA) KFA-Net all show excellent effects. Compared...

Download This Paper Open PDF in Browser Add to My Library Share: Permalink Using these links will ensure access this page indefinitely Copy URL DOI

This letter proposes a novel single image super-resolution (SR) method based on the low-rank matrix recovery (LRMR) and neighbor embedding (NE). LRMR is used to explore underlying structures of subspaces spanned by similar patches. Specifically, training patches are first divided into groups. Then technique utilized learn latent structure each group. The NE algorithm performed learnt components HR LR produce SR results. Experimental results suggest that our approach can reconstruct high...

The haze in remote sensing images can cause the decline of image quality and bring many obstacles to applications images. Considering non-uniform distribution images, we propose a single dehazing method based on encoder–decoder architecture, which combines both wavelet transform deep learning technology. To address clarity issue with haze, preliminary process input by atmospheric scattering model, extract first-order low-frequency sub-band information its 2D stationary as an additional...

Target detection technology has been greatly improved for the synthetic aperture radar (SAR) images recently, due to advancement in deep learning (DL) domain. However, because of existence clutter SAR images, it's still a challenge detect small targets with high accuracy and low computational complexity. To solve this problem, algorithm based on feature fusion cross-layer connection (FFCLC) network is proposed paper. Firstly, attention (AFF) applied improve ability through allocating weights...

Remote sensing images are very vulnerable to cloud interference during the imaging process. Cloud occlusion, especially thick significantly reduces quality of remote images, which in turn affects a variety subsequent tasks using images. The miss ground information due occlusion. removal method based on temporality global–local structure is initially suggested as solution this problem. This includes two stages: global multi-temporal feature fusion (GMFF) stage and local single-temporal...

Idiopathic pulmonary fibrosis (IPF) is a fatal lung disease characterized by fibroblast proliferation and extracellular matrix remodeling; however, the molecular mechanisms underlying its occurrence development are not yet fully understood. Despite it having variety of beneficial pharmacological activities, effects catalpol (CAT), which extracted from Rehmannia glutinosa , in IPF known. In this study, differentially expressed genes, proteins, pathways Gene Expression Omnibus database were...

At the time of prevalence coronavirus disease 2019 (COVID-19), pulmonary fibrosis (PF) related to COVID-19 has become main sequela. However, mechanism PF COVID (COVID-PF) is unknown. This study aimed explore key targets in development COVID-PF and d -limonene treatment. The differentially expressed genes were downloaded from GeneCards database, their pathways analyzed. -Limonene was molecularly docked with proteins screen its pharmacological targets, a rat lung model established verify...

High-quality remote sensing images play important roles in the development of ecological indicators’ mapping, urban-rural management, urban planning, and other fields. Compared with natural images, have more abundant land cover along lower spatial resolutions. Given embedded longitude latitude information reference (Ref) similar scenes could be accessible. However, existing traditional super-resolution (SR) approaches always depend on increases network depth to improve performance, which...

Abstract Capturing the dynamics of urban fire situation is a basic but challenging task, which takes an indispensable role in field security and emergency decision. Traditional methods approach prediction via stochastic process based on physics or statistics, may be interpretable less practical real applications. Recently, some data-driven models, Convolutional Neural Network (CNN), Recurrent (RNN) Graph (GCN), seem to fruitful capturing spatio-temporal with massive high-dimension data. In...

A novel super-resolution (SR) method is proposed in this paper to reconstruct high-resolution (HR) remote sensing images. Different scenes of images have great disparities structural complexity. Nevertheless, most existing SR methods ignore these differences, which increases the difficulty train an network. Therefore, we first propose a preclassification strategy and adopt different networks process with Furthermore, main edge low-resolution are extracted as shallow features fused deep by...

The spatter generated by the interaction between laser and powder during Powder Bed Fusion-Laser Melting (PBF-LM) can significantly affect quality of printed parts. A high-speed camera is used to observe dynamic process spatter’s behavior under different layer thickness powers printing process, analyze samples’ surface roughness, microstructure, mechanical properties. In terms image processing, employing an optical flow approach track quantify number spatters efficiently eliminates...

Low‐resolution (LR) document images may cause difficulties in reading or low recognition rates computer vision. Thus, it is necessary to improve the resolution of an LR image via some algorithms. In this study, a novel super‐resolution (SR) method using structural similarity and Markov random field (MRF) proposed. First, non‐local algorithm utilised find similar patches. Instead Euclidian distance, modified chi‐square distance proposed measure patch because bimodality characteristic can be...

Ship motions are usually employed to analyze water transportation system. Interpolation of ship can be used for estimating the lost movement information, which is important analyzing traffic. Although some methods proposed interpolate trajectory, these needed improve interpolation accuracy. In addition, course and speed not interpolated in previous studies. This article proposes a novel estimation algorithm interpolating motions. First, bilateral filtering smooth navigation trajectories,...

In general, the space-time adaptive processing (STAP) can achieve excellent clutter suppression and moving target detection performance in airborne multiple-input multiple-output (MIMO) radar for increasing system degrees of freedom (DoFs). However, improvement is accompanied by a dramatic increase computational cost training sample requirement. As one most efficient dimension-reduced STAP methods, extended factored approach (EFA) transforms full-dimension problem into several small-scale...

This paper presents a new approach to single-image super-resolution reconstruction, based on sparse signal representation using classified dictionaries. The high-resolution and low-resolution image patches training sets are divided into two categories respectively by classification templates which give consideration direction edge features. Then, we train pair of learning dictionaries illustrate the features both kinds patches. Learning combined with realize reconstruction. As experiment...