计算机视觉(视频追踪检测分类、监控追踪)常用测试数据集
(1)、WallFlower dataset 【链接】:
用于评价背景建模算法的好坏, Ground-truth foreground provided.
(2)、Foreground/Background segmentation and Stereo dataset 【链接】:
from Microsoft Cambridge.
(3)、VISOR: Video Surveillance Online Repositiory 【链接】:
大量的视频和路面实况.
(4)、3D Photography Dataset 【链接】
(5)、Multi-model, multi-camera meeting room dataset 【链接】
(6)、Advanced Video and Signal based Surveillance 【链接】:
各种用于跟踪和检测的数据集
(7)、Caltech image collections 【链接】:
用于目标物体检测,分割和分类
(8)、INRIA Datasets 【链接】:
车辆, 人, 马, 人类行为等
(9)、CAVIAR surveillance Dataset 【链接】
(10)、Videos for Head Tracking 【链接】
(11)、Pedestrian dataset from MIT 【链接】
(12)、Shadow detection datasets 【链接】
(13)、Flash and non-Flash dataset 【链接】
(14)、Experiments on skin region detection and tracking 【链接】:
包括一个ground-truthed dataset
(15)、MIT Face Dataset 【链接】
(16)、MIT Car Datasets 【链接】
(17)、MIT Street Scenes 【链接】:
CBCL StreetScenes Challenge Framework是一个图像、注释、软件和性能检测的对象集[cars, pedestrians, bicycles, buildings, trees, skies, roads, sidewalks, and stores]
(18)、LabelMe Dataset 【链接】:
超过150,000已经标注的照片
(19)、MuHAVi 【链接】:
Multicamera Human Action Video Data,A large body of human action video data using 8 cameras. Includes manually annotated silhouette data. 用于测试人行为的数据集
(20)、INRIA Xmas Motion Acquisition Sequences (IXMAS) 【链接】:
Multiview dataset for view-invariant human action recognition.
(21)、i-LIDS datasets 【链接】:
UK Government benchmark datasets for automated surveillance.
(22)、The Daimler Pedestrian Detection Benchmark 【链接】:
contains 15,560 pedestrian and non-pedestrian samples (image cut-outs) and 6744 additional full images not containing pedestrians for bootstrapping. The test set contains more than 21,790 images with 56,492 pedestrian labels (fully visible or partially occluded), captured from a vehicle in urban traffic.
(23)、Stereo Pedestrian Detection Evaluation Dataset 【链接】:
a dataset for evaluating pedestrian detection using stereo camera images and video. 用于测试行人检测算法的数据集
(24)、Colour video and Thermal infrared datasets 【链接】:
Dataset of videos in colour and thermal infrared. Videos are aligned temporally and spatially. Ground-truth for object tracking is provided.
(25)、北航数字媒体【链接】:
面向室外视频监控的运动目标检测跟踪库
(26)、google结果:
http://www.cvpapers.com/datasets.html
http://riemenschneider.hayko.at/vision/dataset/
http://homepages.inf.ed.ac.uk/cgi/rbf/CVONLINE/entries.pl?TAG363
转自:http://blog.csdn.net/lucky_greenegg/article/details/10241295