The rapid increase in the volume of video data generated from edges Industrial Internet Things, opens up new possibilities for enhancing application service. Multicamera multiobject tracking (MCMT) has always been a fundamental task surveillance or traffic control. However, traditional MCMT methods are limited by communication bottleneck and computation resources centralized curator, suffer security privacy issues. In this article, we first design multicamera multihypothesis (MC-MHT)...

With the development of smart cities, video surveillance has become more prevalent in urban areas. The rapid growth data brings challenges to processing and analysis. Multi-object tracking (MOT), one most fundamental tasks computer vision, a wide range applications prospects. MOT aims locate multiple objects maintain their unique identities by analyzing frame frame. Most existing frameworks are deployed centralized systems, which convenient for management but have problems such as weak...

With the explosive growth of Internet traffic, large sensitive and valuable information is at risk cyber attacks, which are mostly preceded by network reconnaissance. A moving target defense technique called host address mutation (HAM) helps facing However, there still exist several fundamental problems in HAM: 1) current approaches cannot be self-adaptive to adversarial strategies; 2) state time-varying because each decides whether mutate IP address; 3) most methods mainly focus on...

Vehicle Re-Identification (ReID) aims to find images of the same vehicle from different videos. It remains a challenging task in video analysis field due huge appearance discrepancy cross-view matching and subtle difference similar vehicles same-view matching. In this paper, we propose Co-occurrence Attention Net (CAN) deal with these two challenges. Specifically, CAN consists branches, main branch an aware branch. The is charge extracting global features that are consistent most views. This...

Compared to 2D imaging data, the 4D light field (LF) data retains richer scene's structure information, which can significantly improve computer's perception capability, including depth estimation, semantic segmentation, and LF rendering. However, there is a contradiction between spatial angular resolution during image acquisition period. To overcome above problem, researchers have gradually focused on super-resolution (LFSR). In traditional solutions, achieved LFSR based various...

Graph Neural Networks (GNNs) serve as a powerful framework for representation learning on graph-structured data, capturing the information of nodes by recursively aggregating and transforming neighboring nodes' representations. Topology in graph plays an important role representations impacts performance GNNs. However, current methods fail to adequately integrate topological into learning. To better leverage enhance capabilities, we propose Attention (GTAT). Specifically, GTAT first extracts...

Vehicular Ad hoc Networks (VANETs) are prone to packet drop attacks because of their inherent distributed architecture and dynamic topology. Existing security schemes mainly focus on multi-path trust-based routing. Unfortunately, the former causes high energy consumption latter requires trust assessment, which is not easy implement in practice. Route mutation (RM) emerging as an active defense technology that changes routes periodically. Traditional RM conceived for fixed network topologies,...

Quality of Experience (QoE) reflects end users' overall experience and feeling with network services, but needs support in terms end-to-end Service (QoS). Segment routing (SR) as a new paradigm can provide good QoS guarantee, making traditional multimedia traffic more efficient scalable. In this paper, we address two problems related to the SR mechanism: enabling fine-grained under complex environment constructing multicast tree branch node load balancing. To solve these problems, an...

In the current industrial informatics society, numerous cameras deployed in modern city promote development of various video services, such as security monitoring and object retrieval. However, traditional methods encounter data leakage risks. Some camera owners are reluctant to share their since contains confidential information. Meanwhile, domain diversities between bring obstacles practical retrieval applications. To deal with these dilemmas, we propose a blockchain-based collaborative...

Robust scene perception is an essential prerequisite to ensure the reliability in ship autonomous driving. However, it a challenging task inland river because of complicated and changeable environment as well high-density ships narrow waterway. As one primary technologies, obstacle trajectory locating tracking has been widely explored recent years. Current approaches strictly rely on lidar only depth awareness sensor limited measurement range severely restricts them for distant object...

Vehicle re-identification (ReID) is a hot topic in intelligent city surveillance. With the development of smart cameras and vehicular edge computing (VEC), numerous media data has opened up new possibilities for enhancing applications vehicle ReID. However, traditional systems face following challenges: 1) it difficult to recognize identities vehicles various views similar appearance, 2) current system hard be extended large-scale low-trust VEC environment. To solve these problems, we...

Offline reinforcement learning represents a pivotal area of advancement within the broader realm learning. Its central objective is to train an agent exclusively using behavioral data, eliminating any need for online interaction. However, relying solely on insights from offline datasets can often lead ineffective solutions, primarily due mismatching between learned policy's understanding and actual underlying environment. Recent research efforts have tended approach this challenge with...

Abstract The coronavirus pandemic has seriously affected public health and social order. Prediction methods based on machine learning can identify the infectivity phenotype risk of coronavirus. Currently, six types coronaviruses that infect humans have been discovered, with significant differences in viral genome sequences. Continuous genetic variation virus will lead to reduced performance models potential forgetting. To solve this challenge, we propose an incremental knowledge distillation...

The growing multimedia services have brought unprecedented challenges to the traditional static network architecture. Moving Target Defense (MTD) has been proposed solve inherent disadvantages of existing defense techniques. As an important area MTD research, Route Mutation (RM) can dynamically change forwarding routes in network. In our previous work, we applied Reinforcement Learning (RL) RM. However, there are still two problems that need be addressed. 1) We consider too few constraints...

To obtain a picking and planting integrated transplanting mechanism (PITM) with ideal trajectory shape, reasonable operating attitude, compact structure for vegetable pot seedling transplanting, symmetrical structural PITM composed of two planetary carriers driven by cam–noncircular gear combination was proposed in this paper. In accordance the requirements agronomy, shape motion attitude end tip clamping claw at several specific positions (the starting points seedling, point pushing special...

This paper aims to introduce a method of roadway and road markings extraction using mobile laser scanning (MLS) data solve the problems inefficiency incompleteness with existing methods. Firstly, spatially reconstructs roadway, extracts ground point cloud by Euclidean clustering, filters out vehicle noises based on K-Means++ clustering algorithm. Then, framework marking semantic identification is developed. The corrected surface further segmented improved dynamic threshold method. minimum...

This paper studies the motion planning problem of pick-and-place an aerial manipulator that consists a quadcopter flying base and Delta arm. We propose novel partially decoupled framework to solve this problem. Compared state-of-the-art approaches, proposed one has two features. First, it does not suffer from increased computation in high-dimensional configuration spaces. That is because calculates trajectories end-effector separately Cartesian space based on geometric feasibility...

This paper proposes a high-fidelity simulation framework that can estimate the potential safety benefits of vehicle-to-infrastructure (V2I) pedestrian strategies. simulator support cooperative perception algorithms in loop by simulating environmental conditions, traffic and characteristics at same time. Besides, benefit estimation model applied our systematically quantify both risk conflict (non-crash condition) severity pedestrian's injuries (crash condition). An experiment was conducted...