Contexts play an important role in salient object detection (SOD). High-level contexts describe the relations between different parts/objects and thus are helpful for discovering specific locations of objects while low-level could provide fine detail information delineating boundary objects. However, way perceiving/leveraging rich has not been fully investigated by existing SOD works. The common context extraction strategies (<i>e.g.</i>, leveraging convolutions with large kernels or atrous...

Abstract This study examines the impact of Belt and Road Initiative (BRI) on China's export potential at product level. Using country–product‐level data from Centre d'Etudes Prospectives d'Informations Internationales‐Base pour l'Analyse du Commerce International (CEPII‐BACI), which includes 5,049 types across 196 economies 2007 to 2022, it explores trade effects BRI its mechanisms perspective global value chains. The results indicate that has increased significantly participating countires....

The microservice-based smart grid service (SGS) organization and the heterogeneous multi-cores-based computing resource supply are development direction of edge in with high penetration renewable energy sources market-oriented. However, their application also challenges schedule for apparatus (ECA), physical carrier computing. In traditional scheduling strategy SGS, an SGS usually corresponds to independent or component, multi-core environment is not considered, making it difficult cope...

Oxidative stress-induced mitochondrial damage is the major cause of cardiomyocyte dysfunction. Therefore, maintenance function, which regulated by quality control (MQC), necessary for homeostasis. This study aimed to explore underlying mechanisms N-acetylcysteine (NAC) function and its relationship with MQC.

With the popular use of geotagging images, more and research efforts have been placed on geographical scene classification. In classification, valid spatial feature selection can significantly boost final performance. Bag visual words (BoVW) do well in selecting classification; nevertheless, it works effectively only if provided extractor is well-matched. this paper, we convolutional neural networks (CNNs) for optimizing proposed extractor, so that learn suitable vocabularies from images....

The rapid development of hardware technology make it possible for computers or even micro-computers to handle the problem traffic information gathering utilizing image/video processing and pattern recognition. In order measure speed a vehicle, several models are proposed in this paper. Firstly, background subtraction method with proper updating model is used detect moving whose position image then traced obtained distinct features. real vehicle world coordination system corresponding pixel...

The rapid development of hardware technology make it possible for computers or even micro-computers to handle the problem traffic information gathering utilizing image/video processing and pattern recognition. In order measure speed a vehicle, several models are proposed in this paper. Firstly, background subtraction method with proper updating model is used detect moving whose position image then traced obtained distinct features. real vehicle world coordination system corresponding pixel...

To overcome the deficiencies of existing methods used in estimation crowd flow with high-density and multi-motion direction, a method based on dynamic texture generalized regression neural network (GRNN) is presented this paper. The firstly extracts features through optical flow, performs moving segmentation by level set algorithm to achieve ROIs, then analysis GRNN between ROI adopted real-time results scene. Experimental show that proposed more suitable than applications low complexity,...

In this cellular network, Relay Stations (RSs) help to improve the system performance; however, little work has been done considering fairness of RSs. paper, we study cooperative game approaches for scheduling in wireless relay networks with two-virtual-antenna array mode. After defining metric channel capacity, form a and present interpretation three different utilization objectives physically mathematically. Then, Nash Bargaining Solution (NBS) is utilized resource allocation traffic load...

One of tasks for how to reduce fiber optic gyroscope (FOG) random noise and drift is potentially useful improving the precision inertial navigation system (INS). In this paper, a novel signal denoising method proposed that combines detrended fluctuation analysis (DFA) empirical mode decomposition (EMD), named DFA-EMD. The noisy first adaptively broken down into oscillatory components called intrinsic functions (IMFs) by EMD. Then DFA used measure each mode, which independent comparison with...

In an embedded system, the hardware resources are limited. order to obtain such traffic information as volume and vehicle speed in a series of efficient video processing algorithms optimization techniques proposed. The key algorithm detect vehicles is background subtraction method, which approximated median filter applied simulated background. Parallelogrammatic "virtual loop" used restrict calculation interested region scene. Tracking done by detection tracking front it. Results show that...

For a robot used to deliver optical fiber for communication, its center of mass and load change over time. This may affect the robot's obstacle-surmounting capability. In this paper, step-climbing performance articulated tracks robots was discussed, regarding maximum height step in term length releasing that carries swinging angle tracks. Two kinds tracked mobile were studied, which four-track with two arms six-track four arms. The relation equations height, released length, angles swing...

Face detection is an important first step towards solving plethora of other computer vision problems. A fast robust face algorithm under a skin color model with geometry constraints proposed in this paper. An ellipse the YCbCr space used to detect regions. light compensation utilized original image, followed by box hair The detected candidate and regions are then combined faces image. Experiments show that method effective efficient.

Camera model identification has been studied extensively within digital image forensics as a deterrence to secret photography and forgery. The feasibility of convolutional neural networks (CNNs) proven for an classification algorithm. CNN-based algorithms have proposed camera classification, they focus on training with image/noise in spatial domain. However, because the periodic characteristics noise are one essential types information it is more efficient train CNN models frequency In this...

Nonlinear impairments seriously affect the transmission performance of high-speed visible light communication (VLC) systems, which have become bottleneck VLC systems in practical applications. To compensate for nonlinear we propose a Pattern quantization method (PQM) based on Gaussian feature inputs and Bayesian learning time-domain. First, investigate signal distribution presence additive white noise (AWGN) impairments. Second, probability statistics are used to quantify noise. Finally,...

This paper introduces the basic principle of non-negativity and support constraints recursive inverse filtering (NAS-RIF) algorithm shortcoming it. We propose an improved scheme. Firstly we use lifting wavelet de-noising technique to smoothen blurred image, then in each iteration NAS-RIF process adopt image segmentation find out precise constraints. Comparing with traditional methods, this method has a positive improvement restraining noise enlargement, maintaining edge features, improving...

Compared to traditional classroom teaching, the SPOC (Small Private Online Course) English teaching model offers greater flexibility but also presents certain challenges. This study proposes optimization of through application personalized autonomous learning theory.

Traditional hand-craft designed methods to accelerated T1rho mapping have limited characterisation capabilities, while deep learning lack the interpretability. On other hand, joint distribution is most direct and accurate way characterize correlation between different images. Therefore, we attempt propose a estimation method use it construct T1$$$\rho$$$ reconstruction model. In particular, score-based diffusion model of acquired T1rho-weighted Moreover, corresponding solved using Langevin...