Crowdsourcing is a process of acquisition, integration, and analysis big heterogeneous data generated by diversity sources in urban spaces, such as sensors, devices, vehicles, buildings, human. Especially, nowadays, no countries, communities, person are immune to emergency events. Detection about events, e.g., fires, storms, traffic jams great importance protect the security humans. Recently, social media feeds rapidly emerging novel platform for providing dissemination information that...

Recent research shows that multimedia resources in the wild are growing at a staggering rate. The rapid increase number of has brought an urgent need to develop intelligent methods organize and process them. In this paper, semantic link network model is used for organizing resources. A whole generating association relation between using proposed. definitions, modules, mechanisms proposed method. integration provides new prospect them with their semantics. tags surrounding texts measure...

Color, as a notable and stable attribute of vehicles, can serve useful reliable cue in variety applications intelligent transportation systems. Therefore, vehicle color recognition natural scenes has become an important research topic this area. In paper, we propose deep-learning-based algorithm for automatic recognition. Different from conventional methods, which usually adopt manually designed features, the proposed is able to adaptively learn representation that more effective task...

Summary An emergency event is an unexceptional that exceeds the capacity of normal resources and organization to cope a situation poses immediate risk health, life, property, or environment. Crowdsourcing connects unobtrusive ubiquitous sensing technologies, advanced data management analytics models, novel visualization methods, create solutions improve urban environment, human life quality, city operation systems. The crowdsourcing on social media can be used detect analyze events. In this...

Temporal grounding aims to localize temporal boundaries within untrimmed videos by language queries, but it faces the challenge of two types inevitable human uncertainties: query uncertainty and label uncertainty. The uncertainties stem from subjectivity, leading limited generalization ability grounding. In this work, we propose a novel DeNet (Decoupling Debias) embrace uncertainty: Decoupling — We explicitly disentangle each into relation feature modified feature. feature, which is mainly...

Gait recognition is a biometric technology that distinguishes individuals by analyzing their walking patterns and holds significant potential for development. Current methods primarily focus on extracting gait features either from the overall appearance or specific local regions. However, they often overlook partitioning of regions based individual's body structure, as well weighted relationships between global features. We propose Weighted Global-Local Feature Fusion Module to partition...

Topology control is an effective method to improve the energy-efficiency and increase communication capacity of Wireless Sensor Networks (WSNs). Traditional topology algorithms are based on deterministic model that fails consider lossy links which provide only probabilistic connectivity. Noticing this fact, we propose a novel network model. We meter connectivity using reachability. It defined as minimal upper limit end-to-end delivery ratio between any pair nodes in network. attempt find...

Hashing is a widely adopted method based on an approximate nearest neighbor search and used in large-scale image retrieval tasks. Conventional learning-based hashing algorithms employ end-to-end representation learning, which one-off technique. Because of the tradeoff between efficiency performance, conventional methods must sacrifice code length to improve increases their computational complexity. To binary codes, motivated by "nonsalient-to-salient" attention scheme humans, we propose...

In this paper, we study the problem of recognizing car brands in surveillance videos, cast it as an image classification problem, and propose a novel multiple instance learning method, named Spatially Coherent Discriminative Pattern Learning, to discover most discriminative patterns images. The learned can effectively distinguish cars different with high accuracy efficiency. experimental results demonstrate that our method is significantly superior recent methods on problem. proposed able...

An emergency event is a sudden, urgent, usually unexpected incident or occurrence that requires an immediate reaction assistance for situations, which plays increasingly important role in the global economy and our daily lives. Recently, web becoming information provider repository due to its real-time, open, dynamic features. In this paper, resources based states detecting algorithm of developed order let people know clearly help social group government process events effectively. The...

Abstract The challenge of multi-object tracking stands as a fundamental focus in computer vision research, finding widespread applications areas such public safety, transportation, autonomous vehicles, robotics, and other domains involving artificial intelligence. Given the intricate nature natural scenes, occurrence object occlusion semi-occlusion is commonplace basic tasks. These factors often result challenges ID switching, loss, detection errors, misaligned bounding boxes, thereby...

Association link network (ALN) can organize massive Web information to provide many intelligent services in our big data society. Effective semantic layered technologies not only theoretical support for knowledge discovery resources, but also improve the searching efficiency of related systems such as system and industrial system. How realize layer division association by hierarchy analysis ALN is an important research topic. To solve this problem, paper proposes a hierarchy-cutting model...