CVRR dataset for trajectory clustering

Six trajectory clustering datasets (Morris and Trivedi, 2009) are provided for benchmarking trajectory clustering algorithms. These datasets cover a diverse range of scenes to enable thorough evaluation of different algorithms. The datasets include three simulated highway scenes, a real highway scene, a simulated intersection scene, and an indoor omnidirectional camera scene.These datasets originate from the work by Morris and Trivedi (2009), described in their paper "Learning trajectory patterns by clustering: Experimental studies and comparative evaluation," presented at the 2009 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) Workshops. The datasets provide a valuable resource for the trajectory clustering community, enabling comparative evaluation and benchmarking of various clustering algorithms. The datasets can be accessed through the following DOI link: https://doi.org/10.1109/CVPR.2009.5206559.

本数据集包含6个源自Morris与Trivedi（2009）的轨迹聚类（trajectory clustering）数据集，用于轨迹聚类算法的基准测试。这些数据集涵盖多样化应用场景，可支持对各类轨迹聚类算法进行全面评估。数据集包含3个模拟高速公路场景、1个真实高速公路场景、1个模拟交叉路口场景以及1个室内全向相机场景。本数据集源自Morris与Trivedi 2009年的研究工作，相关成果发表于他们在2009年IEEE计算机视觉与模式识别会议（CVPR）研讨会上提交的论文《基于聚类的轨迹模式学习：实验研究与对比评估》。本数据集为轨迹聚类领域提供了宝贵的研究资源，可支撑各类聚类算法的对比评估与基准测试工作。用户可通过以下DOI链接获取本数据集：https://doi.org/10.1109/CVPR.2009.5206559

搜集汇总

背景与挑战

背景概述

CVRR数据集包含六个轨迹聚类数据集，涵盖多种场景（模拟和真实高速公路、交叉路口及室内场景），源自Morris和Trivedi（2009）的研究，用于轨迹聚类算法的基准测试和比较评估。数据集大小为2.29 MB，采用CC BY 4.0许可。