A5050231987- Evacuation and Crowd Dynamics
- Video Surveillance and Tracking Methods
- Differential Equations and Boundary Problems
- Image Retrieval and Classification Techniques
- Traffic and Road Safety
- Facility Location and Emergency Management
- Robotic Path Planning Algorithms
- Advanced Image and Video Retrieval Techniques
- Flood Risk Assessment and Management
- Advanced Measurement and Detection Methods
- 3D Shape Modeling and Analysis
- Advanced Vision and Imaging
- Numerical methods for differential equations
- Anomaly Detection Techniques and Applications
- 3D Surveying and Cultural Heritage
- Educational Technology and Assessment
- Differential Equations and Numerical Methods
- Simulation and Modeling Applications
- Soil, Finite Element Methods
- Nonlinear Differential Equations Analysis
Anhui Normal University
2019-2023
Hengshui University
2004-2010
With the rapid development of Web Virtual Reality (WebVR) technology, increasing focus has been placed on this domain. WebVR indoor scenario design studies have important value in both academia and industry. However, many bottlenecks still need to be overcome, such as weak computing capacity limited memory space available web browsers. In particular, there are deficiencies virtual scenes, i.e., lack fidelity, low automation poor scene interactability. paper, we propose a novel furniture...
In this article, we present a deep learning approach to sketch-based shape retrieval that incorporates few novel techniques improve the quality of results. First, address problem scarcity training sketch data, augmentation method more closely mimics human sketches compared simple image transformation. Our generates from existing data by (i) removing stroke, (ii) adjusting and (iii) rotating sketch. As such, generate large number samples for our neural network. Second, obtain 2D renderings...
Abstract One of the most common solutions for prevention fire accidents is to conduct extensive evacuation drills in crowded places. However, there are multiple salient advantages using virtual reality technology simulate emergency solutions, instance, saving costs and greatly decreasing uncertain risks or accidents. Therefore, this article, a multiagent framework complex scenarios proposed effectively used procedure approximate goal less costly manner. Specifically, concept multihierarchy...
With the rapid development of Web3D, virtual reality, and digital twins, trajectories decision data considerably rely on analysis understanding real video data, particularly in emergency evacuation scenarios. Correctly effectively evacuating crowds scenarios are becoming increasingly urgent. One good solution is to extract pedestrian from videos situations using a multi-target tracking algorithm use them define procedures. To implement this solution, trajectory extraction optimization...
目的 随着虚拟现实技术的发展，在虚拟场景中，基于多智能体的逃生路径规划已成为关键技术之一。与传统的火灾演习相比，采用基于虚拟现实的方法完成火灾逃生演练具有诸多优势，如成本低、代价小、可靠性高等，但仍有一定的局限性，为此，提出一种改进的双层深度Q网络（deep Q network，DQN）架构的路径规划算法。方法 基于两个结构相同的双Q网络，优化了经验池的生成方法和探索策略，并在奖励中增加火灾这样的环境因素对智能体的影响。同时，为了提高疏散的安全性和效率，提出了一种基于改进的K-medoids算法的多智能体分组策略方法。结果 相关实验表明提出的改进的双层深度Q网络架构收敛速度更快，学习更加稳定，模型性能得到有效提升。综合考虑火灾场景下智能体的疏散效率和疏散安全性，使用指标平均健康疏散值（average health evacuationvalue，AHEP）评估疏散效果，相较于传统的路径规划方法A-STAR （a star search algorithm）和DIJKSTRA （Dijkstra’...