Vehicle re-identification (ReID) technology has played a more and important role in Intelligent Transport System (ITS), which aims at searching the same query vehicle identity from large amount of gallery datasets under different non-overlapping camera views. Current related researches mainly focus on discriminative feature mining images train model fully supervised manner highly relies manual annotations training data. However, it is labor-consuming impractical to generate annotation for...

Person re-Identification (reID), aiming at retrieving a person across different cameras, has been playing more and important role in the construction of smart city social security. For deep-learning-based reID methods, it proved that using local feature together with global could help to give robust representation for retrieval. Human pose information can provide locations human skeleton effectively guide network pay attention these key areas, also reduce noise distractions from background...

While speed to market is a critical success for outsourcing software projects, the quality of increasing importance both customer and development teams. When legacy system reengineering project, requirement generation, business confirmation, experts finding should be well planned, organized performed in order achieve high quality. In lattice extreme programming concepts were used. Iterative reduced impact communication delay while there always enough tasks do would not block on waiting...

Concept lattice is an effective tool for knowledge representation and data analysis. It have been successfully applied to many fields. This paper illustrates the theory of compact dependencies in concept lattice, then study approach briefest rules extraction based on dependencies, including decision association with confidence 1. Finally, method example relationship between process parameters aim parameters,as well as other parameters, are discussed. provides a useful support advanced...

The clonal selection mechanism and vaccination strategy of immune system are introduced into particle swarm optimization algorithm in this paper, order to enhance the ability global exploration PSO, avoiding getting local optimum improving accuracy convergence speed BP networks.The Cauchy mutation operator Gauss used improve searching optimization.Then weights thresholds neural networks trained by applying optimization.Finally coke ratio forecasting model is established based on modified...

The Iterative Closest Point(ICP) algorithm is usually used for 3D ear recognition in the literatures. However, high computational cost of ICP limits application biometrics. In this paper, we present an efficient approach based on local feature and recognition. features are detected represented with LSP(Local Surface Patch), to compute initial transformation matching ICP. An elite preservation strategy introduced refine candidate gallery ears that a modified kd-tree index, distance uniqueness...

(1) School of Automation and Electrical Engineering, University Science Technology Beijing, China; (2) China Nuclear Power Engineering Co.,Ltd,

Legacy software systems present a real challenge to the people who use and maintain them. Dealing with these decayed but still working is also big problem. To tackle it, many owners choose re-engineer Obviously, new system's maintainability one of most important issues considered in re-engineering project. In last few years, methodologies technologies have been studied achieve successful reengineering process, formal concepts analysis (FCA) used analyze speed up This paper proposes an...

Video-based person re-identification (Re-ID) has drawn more attention as video surveillance could offer richer spatial and temporal information to potentially reduce visual ambiguities occlusion. For ambiguities, multi-scale features are beneficial for distinguishing similar pedestrian sequences by different semantic information. occlusion, Graph Convolutional Network (GCN) effectively leverage the complementary with node pairs Re- ID task. In this paper, we propose a novel Multi-Scale...

Person re-identification (reID) aims at retrieving a person from images captured by different cameras. For deep-learning-based reID methods, it has been proved that using local features together with global feature could help to give robust representation for retrieval. Human pose information provide the locations of human skeleton effectively guide network pay more attention on these key areas and also reduce noise distractions background or occlusion. However, methods proposed previous...

This paper proposes a novel baseline for deep person ReID methods by introducing human pose-based attention mechanism. Benefiting from convolutional network, there has been great progress of re-identification (ReID) in recent years, which aims at retrieving the same identities images captured different cameras. Most existing focus on designing complex network structures to achieve higher scores public datasets, but few works pay design. A strong is crucial experiments and could make...

In recent years, there has been a great progress on object-tracking task. Most of them are based Joint Detection and Embedding (JDE) benchmark, which accomplish detection Re-ID task in single module thus could reduce the time cost help to gain higher processing FPS. However, large computation requirement existing JDE-based method, usually demand several expensive GPU devices, is still an obstacle wide application for industry. this paper, we propose new lightweight structure named...