With the ever-increasing data processing capabilities of edge computing devices and growing acceptance running social sensing applications on such cloud-edge systems, effectively allocating tasks between server has emerged as a critical undertaking for maximizing performance systems. Task allocation in an environment faces several unique challenges: (i) objectives may be inconsistent or even conflicting with each other, (ii) only partially collaborative finishing computation due to "rational...

The object detection based on deep learning is an important application in technology, which characterized by its strong capability of feature and representation compared with the traditional methods. paper first makes introduction classical methods detection, expounds relation difference between detection. Then it introduces emergence elaborates most typical nowadays via learning. In statement methods, focuses framework design working principle models analyzes model performance real-time...

With the rapid growth of online social media and ubiquitous Internet connectivity, sensing has emerged as a new crowdsourcing application paradigm collecting observations (often called claims) about physical environment from humans or devices on their behalf. A fundamental problem in applications lies effectively ascertaining correctness claims reliability data sources without knowing either them priori, which is referred to truth discovery. While significant progress been made solve...

A multi-view object detection approach based on deep learning is proposed in this paper. Classical methods regression models are introduced, and the reasons for their weak ability to detect small objects analyzed. To improve performance of these methods, a proposed, model structure working principles explained. Additionally, retrieval accuracy both corresponding classical evaluated compared test dataset. The experimental results show that terms capability, Multi-view YOLO (You Only Look...

Container management frameworks, such as Docker, package diverse applications and their complex dependencies in self-contained images, which facilitates application deployment, distribution, sharing. Currently, Docker employs a shared-nothing storage architecture, i.e. every Docker-enabled host requires its own copy of an image on local to create run containers. This greatly inflates utilization, network load, job completion times the cluster. In this paper, we investigate option storing...

Workflows are a widely used abstraction for describing large scientific applications and running them on distributed systems. However, most workflow systems have been silent the question of what execution environment each task in is expected to run in. Consequently, may successfully it was created, but fail other platforms due differences environment. Container-based schedulers recently arisen as potential solution this problem, adopting containers distribute computing resources deliver...

Current cytotoxic T lymphocyte (CTL) activating immunotherapy requires a major histocompatibility complex I (MHC-I)-mediated presentation of tumor-associated antigens, which malfunctions in around half patients with triple-negative breast cancer (TNBC). Here, we create LCL161-loaded macrophage membrane decorated nanoparticle (LMN) for MHC-I-deficient TNBC. SIRPα on the helps LMNs recognize CD47-expressing cells targeted delivery LCL161, induces release high mobility group protein 1 and...

Federated learning decentralizes the process, yet it does not provide adequate privacy protection. Current countermeasures predominantly rely on Local Differential Privacy (LDP) techniques. While larger noise injection offers stronger privacy, also leads to a degradation in model performance and necessitates additional training iterations. Existing methods struggle strike balance between user usability, efficiency. To address this, we propose an adaptive scaling method grounded mutual...

This article [J. Electron. Imaging 25(6), 061602 (2016), doi: 10.1117/1.JEI.25.6.061602] was retracted on 18 December 2018 due to double publication in this and another peer-reviewed journal. The authors regret mistake.

High-throughput computing (HTC) workloads seek to complete as many jobs possible over a long period of time. Such require efficient execution parallel and can occupy large number resources for As result, full utilization is the normal state an HTC facility. The widespread use container orchestrators eases deployment frameworks across different platforms, which also provides opportunity scale up with almost infinite on public cloud. However, autoscaling mechanisms are primarily designed...

As traditional visual tracking algorithms based on co-training framework are characterized by poor robustness in numerous real-time scenarios, a robust algorithm structural multi-scale features adaptive fusion is thereby proposed. In order to achieve comprehensive representation, we select gray intensity space and local binary pattern (LBP) respectively train the corresponding Bayes classifiers. Online feature selection criterion via maximizing posterior probability entropy function utilized...

Object tracking is a core subject in computer vision and has significant meaning both theory practice. We propose method which robust discriminative classifier built based on object context information. In this method, we consider multiple frames of local invariant features around the construct template template. To overcome limitation representations, also design nonparametric learning algorithm using transitive matching perspective transformation. This can keep adding appearance avoid...

Federated Learning (FL) is a novel machine learning paradigm that enables multiple participants to collaboratively train model by aggregating local gradients from each client without sharing sensitive data with other. The clients' plaintext transmission makes FL systems vulnerable inference attacks, which aim infer their updates. Masking additive homomorphic encryption (especially the Paillier scheme) before aggregate straightforward way ensure security. Unfortunately, this method needs...

How to accurately estimate facial age is a difficult problem due insufficiency of training data. In this paper, an effective approach proposed by means extreme learning machine (ELM). the method, set features randomly selected from original consist feature subspace. Given initial weight matrix, samples within subspace are input ELM constitute weaker estimator. Besides subspace, matrix varied construct multiple estimators with good diversity. order alleviate negative affect caused sample...

Recently, tracking methods based on discriminative correlation filters (DCF) and CNN features have seen a great improvement in accuracy performance. However, increasingly complex models heavy computation burdens can reduce their speed real-time capability. In this paper, we analyze the key factors that increased state-of-the-art DCF trackers provide some solutions for solving these problems. We propose novel method keypoint consensus clustering improved DCFs. First, use consensus-based...

Object tracking is a core subject in computer vision and has significant meaning both theory practice. In this paper, we propose novel method, which robust discriminative classifier built basing on object context information. consider multiple frames of local invariant features around the object, construct template template. To overcome limitation representations, also design non-parametric learning algorithm using transitive matching perspective transformation, called as LUPT (Learning...

Fault inspection plays an important role in ensuring the safe operation of freight cars. With development computer vision technology, vision-based fault has become one principal means inspection. A coupler yoke is component train’s connection system, and if bolt goes missing, it would cause separation train from coupler, resulting a serious accident. In this paper, we propose automatic image system to inspect faults yokes during train. Images are acquired process divided into two parts:...

The paper proposes a visual tracking method based on deep learning without online model updating. In consideration of the advantages in feature representation, SSD (Single Shot Multibox Detector) is used as object extractor model. Simultaneously, color histogram and HOG (Histogram Oriented Gradient) are combined to select object. process tracking, multi-scale searching map built improve detection performance efficiency. experiment eight respective video sequences baseline dataset, compared...

Objective: A novel structure based on channel-wise attention mechanism is presented in this paper. Embedding with the proposed structure, an efficient classification model that accepts multi-lead electrocardiogram (ECG) as input constructed. Methods: One-dimensional convolutional neural networks (CNN) have proven to be effective pervasive tasks, enabling automatic extraction of features while classifying targets. We implement Residual connection and design a which can learn weights from...

To overcome the drawback that Boosting decision trees perform fast speed in test time while training process is relatively too slow to meet requirements of applications with real-time learning, we propose a method by pruning those noneffective features advance. And basing on this method, also design algorithm. Firstly, analyze structure each node, and prove classification error node has bound through derivation. Then, using boundary prune non-effective early stage, greatly accelerate tree...

Virtual interaction can provide vivid scenarios for the audience, and its operability greatly shortens distance between works audience. The key to building a complete set of virtual is design aesthetic 3D video logical design. Micro plastic pollution an environmental problem that has recently attracted much attention. plastics widely exist in various media carriers all over world. As far as straws are concerned, on average, more than 30 used each year China. A large number micro produced...

Low-resolution Chinese character recognition of license plate is always a difficult problem. For solving it, we must think about the distinctiveness feature and counting speed method simultaneously. In this paper, proposed simple effective extraction algorithm. First, extract statistical based on decomposing stroke with wavelet transform. Second, apply Elastic Mesh Algorithm into extracting coefficient to get structure information character. The experimental results show robust against low...