paper 28 ：一些常见常用数据库的下载网站集锦

做图像处理+模式识别的童鞋怎么可以没有数据库呢？ 但是，如果自己做一个数据库，费时费力费钱先不说，关键是建立的数据库的公信力一般不会高，做出的算法也别人也不好比较，所以呢，下载比较权威的公共数据库还是来得方便、直接、有效。

原先我也收藏了一些公共数据库的下载链接，但是由于近期电脑损坏给搞没了，那个后悔那........ 现在专门收集的这些链接一定要保存好，希望在与大家共同分享的同时，也能收藏些其它新的链接，呵呵。

闲话少说，现在开始：

1， http://www.multitel.be/cantata/

这个网址提供了大量的视频和图像的数据库下载索引，并有相应的介绍，强烈推荐！ 大家慢慢去找寻自己的惊喜吧

2，http://www.cvpapers.com/datasets.html

CV Datasets on the web ， 主要好像是直立行人检测....

3， http://www.cvc.uab.es/adas/site/?q=node/7

里面又有好几种数据库可以下载：CVC Virtual Pedestrian Dataset、CVC-01 Pedestrian Dataset、CVC-02 Pedestrian Dataset

4， http://www.cis.upenn.edu/~jshi/ped_html/

Database description：This is an image database containing images that are used for pedestrian detection in the experiments reported in[1]. The images are taken from scenes around campus and urban street. The objects we are interested in these images are pedestrians. Each image will have at least one pedestrian in it.

The heights of labeled pedestrians in this database fall into [180,390] pixels. All labeled pedestrians are straight up.

There are 170 images with 345 labeled pedestrians, among which 96 images are taken from around University of Pennsylvania, and other 74 are taken from around Fudan University.

5，http://www.vision.caltech.edu/Image_Datasets/CaltechPedestrians/

Caltech Pedestrian Detection Benchmark：The Caltech Pedestrian Dataset consists of approximately 10 hours of 640x480 30Hz video taken from a vehicle driving through regular traffic in an urban environment. About 250,000 frames (in 137 approximately minute long segments) with a total of 350,000 bounding boxes and 2300 unique pedestrians were annotated. The annotation includes temporal correspondence between bounding boxes and detailed occlusion labels. More information can be found in our PAMI 2011 and CVPR 2009 benchmarking papers.

6， http://www.edgar-seemann.de/pd/datasets.py

Pedestrian Detection

7， http://www.wisdom.weizmann.ac.il/~vision/SpaceTimeActions.html#Database  Weizmann 人体行为库

http://www.nada.kth.se/cvap/actions/  KTH人体行为数据库

http://4drepository.inrialpes.fr/public/viewgroup/6   INRIA XMAX多视角视频库

http://vision.eecs.ucf.edu/data.html  UCF Sports 数据库

http://www.di.ens.fr/~laptev/actions/hollywood2/ Hollywood 人体行为库

http://vision.stanford.edu/Datasets/OlympicSports/ Olympic sports dataset

这几个数据库均是基于动作/行为识别的(在第1条网址中也可以找到它们的下载地址)，文章《视频中行为识别公开数据库汇总

》对它们的评价比较中肯，可以参看：http://blog.sina.com.cn/s/blog_631a4cc40101138j.html

8， http://homepages.inf.ed.ac.uk/rbf/BEHAVE/

Computer-assisted prescreening of video streams for unusual activities

9， http://www.cc.gatech.edu/cpl/projects/monsoon/PropagationNet/PropagationNet.htm

Propagation Networks for Recognizing Partially Ordered Sequential Activity. Goals：

	Represent and fuse human knowledge of daily activities with noisy perceptual features
	Detect and recognize an activity
	Pinpoint components of the activity and detect missing or improperly performed steps

10， http://root.simpleinfo.net/1984DA173065/AreaDatum.aspx

由模式识别国家重点实验室提供的链接，数据量比较大，通常需要签属协议，以光盘形式拿到数据。可以下载的有虹膜库数据、掌纹数据库、步态数据库、中文语言资源库、笔迹数据库、三维人脸数据库、行为分析数据库

11，http://www.datatang.com