CG-CrowdSim, CG-BioCrowds
收藏arXiv2019-05-01 更新2024-08-06 收录
下载链接：
http://arxiv.org/abs/1905.00261v1
下载链接
链接失效反馈官方服务：
资源简介：
本研究涉及的数据集包括CG-CrowdSim和CG-BioCrowds，由巴西波尔图阿尔格雷桥天主教大学的技术学院创建。CG-CrowdSim基于UCSD数据集，通过模拟生成20个虚拟人类图像，而CG-BioCrowds则基于学生数据集，模拟了22个虚拟人类。这些数据集通过Unity引擎实时渲染，每秒生成30张图像，用于机器学习方法的训练。数据集的创建旨在通过自动生成带有地面实况的计算机图形图像，简化人群检测和计数的数据集构建过程，从而优化事件检测的整个流程。
The datasets involved in this study are CG-CrowdSim and CG-BioCrowds, which were developed by the Faculty of Technology at the Pontifical Catholic University of Rio Grande do Sul (PUCRS) in Porto Alegre, Brazil. CG-CrowdSim is based on the UCSD dataset, generating 20 virtual human images via simulation. In contrast, CG-BioCrowds is built upon the student dataset and simulates 22 virtual humans. All datasets are rendered in real time using the Unity engine, producing 30 images per second, and are used for training machine learning methods. The purpose of creating these datasets is to simplify the dataset construction process for crowd detection and counting by automatically generating computer graphics images with ground truth, thereby optimizing the entire workflow of event detection.
提供机构：
数据可视化与交互实验室，虚拟人类模拟实验室，技术学院，PUCRS-波尔图阿尔格雷桥天主教大学，阿雷格里港，巴西
创建时间：
2019-05-01
