计算机视觉数据集

转载自： https://blog.csdn.net/lcj_cjfykx/article/details/21217551

http://host.robots.ox.ac.uk/pascal/VOC/voc2012/VOCtrainval_11-May-2012.tar ：VOC2012含多种样本

Pedestrian dataset from MIT： 行人图片背景简单

INRIA 数据集 ：行人检测图片，背景复杂

WallFlower dataset: 用于评价背景建模算法的好坏. Ground-truth foreground provided.
Foreground/Background segmentation and Stereo dataset: from Microsoft Cambridge.
VISOR: Video Surveillance Online Repositiory: 大量的视频和路面实况.
3D Photography Dataset
Multi-model, multi-camera meeting room dataset
Advanced Video and Signal based Surveillance: 各种用于跟踪和检测的数据集.
Caltech image collections: 用于目标物体检测，分割和分类
INRIA Datasets: 车辆, 人, 马, 人类行为等

CAVIAR surveillance Dataset
Videos for Head Tracking
Pedestrian dataset from MIT
Shadow detection datasets
Flash and non-Flash dataset
Experiments on skin region detection and tracking: 包括一个ground-truthed dataset

MIT Face Dataset
MIT Car Datasets
MIT Street Scenes: 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]
LabelMe Dataset: 超过150,000已经标注的照片.
MuHAVi: Multicamera Human Action Video DataA 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
INRIA Xmas Motion Acquisition Sequences (IXMAS): Multiview dataset for view-invariant human action recognition.
i-LIDS datasets: UK Government benchmark datasets for automated surveillance.
The Daimler Pedestrian Detection Benchmark: 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.
Stereo Pedestrian Detection Evaluation Dataset: 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
Colour video and Thermal infrared datasets: Dataset of videos in colour and thermal infrared. Videos are aligned temporally and spatially. Ground-truth for object tracking is provided.