Datasets
BEHAVE Action Dataset full annotationBEHAVE Interactions Test Case Scenarios was presented in 2009 in the paper "The BEHAVE video dataset: ground truthed video for multi-person behavior classification" by S. Blunsden and R.B. Fisher from the University of Edinburgh. It consists in a 1 hour long surveillance video with actors performing different collective actions. The database was aimed as a benchmarking groundtruth for people interaction detection/classification algorithms. During my research, I found BEHAVE dataset very convenient also for forensic video analysis given that it contains different people reappearing and performing different actions through a relatively long video.One of the problems of this dataset is that the annotations correspond to a small part of the video, and are only focused on the interaction itself. Here I provide the persons annotations in the whole video together (5 re-appearing actors plus other not re-appearing not-actors) with the individual actions (walking , fighting, running and standing). The annotations are provided in a single txt file and the format is as follows:
References to these annotations should be made to the following article:
|
|||
CVC-02 Pedestrian DatasetCVC-02 consists of three subsets, each one focused on a different task of pedestrian detection: candidate generation, classification and system evaluation. The imagery has been recorded in urban scenarios around Barcelona (Spain), using a Bumblebee color stereo camera with resolution 640x480 pixels and 6mm focal length. The annotated pedestrians are in the range from 0 to 50 m from the camera, which corresponds to a smallest pedestrian of 12x24 pixels. The main features of each subset are the following:
References to this pedestrian database should be made to the following article:
Download: CVC-02-CandidateGeneration (269MB) Download: CVC-02-Classification (1.5GB) Download: CVC-02-System Sequence 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 (around 1GB each) Please, read the terms of use before downloading the database: disclaimer. |
|||
CVC-01 Pedestrian DatasetThis pedestrian database was recorded in the surreoundings of Barcelona and annotated by the CVC ADAS group in order to evaluate the pedestrian detection algorithms developed. It was originally named CVC-CER-01. Its main features are the following:
References to this pedestrian database should be made to the following article:
Please, read the terms of use before downloading the database: disclaimer. |