Procedural Human Action Videos - Textual Annotations
Cesar Roberto de Souza and Adrien Gaidon and Yohann Cabon and Antonio Manuel Lopez Pena

phav_v2-info (318 files)
info_5009.tar.bz2 57.43MB
info_5006.tar.bz2 54.81MB
info_5007.tar.bz2 58.20MB
info_5008.tar.bz2 58.64MB
info_5003.tar.bz2 57.99MB
info_5004.tar.bz2 60.43MB
info_5005.tar.bz2 60.54MB
info_5000.tar.bz2 57.52MB
info_5001.tar.bz2 54.24MB
info_5002.tar.bz2 59.17MB
info_4021.tar.bz2 83.48MB
info_4022.tar.bz2 48.32MB
info_4023.tar.bz2 52.04MB
info_4018.tar.bz2 61.03MB
info_4019.tar.bz2 56.50MB
info_4020.tar.bz2 57.63MB
info_4015.tar.bz2 54.67MB
info_4016.tar.bz2 57.63MB
info_4017.tar.bz2 56.40MB
info_4012.tar.bz2 55.92MB
info_4013.tar.bz2 55.92MB
info_4014.tar.bz2 54.61MB
info_4009.tar.bz2 58.62MB
info_4010.tar.bz2 57.61MB
info_4011.tar.bz2 49.80MB
info_4006.tar.bz2 60.12MB
info_4007.tar.bz2 56.95MB
info_4008.tar.bz2 54.30MB
info_3122.tar.bz2 61.63MB
info_4005.tar.bz2 54.70MB
info_3119.tar.bz2 59.86MB
info_3120.tar.bz2 60.53MB
info_3121.tar.bz2 58.58MB
info_3117.tar.bz2 59.07MB
info_3118.tar.bz2 58.78MB
info_3115.tar.bz2 53.54MB
info_3116.tar.bz2 57.36MB
info_3112.tar.bz2 61.27MB
info_3113.tar.bz2 58.79MB
info_3114.tar.bz2 61.19MB
info_3109.tar.bz2 56.85MB
info_3110.tar.bz2 55.13MB
info_3111.tar.bz2 59.75MB
info_3106.tar.bz2 54.13MB
info_3107.tar.bz2 56.08MB
info_3108.tar.bz2 58.59MB
info_3103.tar.bz2 57.62MB
info_3104.tar.bz2 58.32MB
info_3105.tar.bz2 61.38MB
Too many files! Click here to view them all.
Type: Dataset
Tags: deep learning, action recognition, procedural generation, video classification

title= {Procedural Human Action Videos - Textual Annotations},
keywords= {Action recognition, procedural generation, deep learning, video classification},
journal= {},
author= {Cesar Roberto de Souza and Adrien Gaidon and Yohann Cabon and Antonio Manuel Lopez Pena},
year= {2017},
url= {},
license= {Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International (CC BY-NC-SA 4.0)},
abstract= { Procedural Human Action Videos (PHAV)
 This torrent contains the Procedural Human Action Videos (PHAV) dataset. This 
 version of the dataset contains 39,982 videos that have been generated by a 
 computer. The dataset is made available under the Creative Commons BY-NC-SA
 (Attribution-NonCommercial-ShareAlike 4.0 International) license. This sub-
 directory contains textual annotations about the contents of these videos,

   - extrinsic.txt: Extrinsic camera parameters
   - 3dpose.txt:    Actor location in camera coordinates
   - bbox2d.txt:    2D bounding boxes in screen coordinates
   - bbox3d.txt:    3D bounding boxes in world coordinates
   - colors.txt:    Color palette for pixel classes in semantic segmentation
   - instances.txt: Color palette for pixel classes in instance segmentation
   - general.txt:   Intrinsic camera parameters and other general information
   - joints.txt:    Body joint locations in screen coordinates (pose)
   - muscles.txt:   Physical properties of body muscles for the main actor
 A human-readable version of the license terms is available at:
 The full license terms are available at:
 By using PHAV, you agree to abide to the terms of the Creative Commons BY-NC-SA
 license. If you do not agree with the terms of this license, please contact the
 dataset authors for other licensing options.
 For more information, including details about how to parse the different files
 included in this release, how to use them in your experiments, and how to cite
 the PHAV dataset in your academical paper, please visit:

superseded= {},
terms= {By using PHAV, you agree to abide to the terms of the Creative Commons BY-NC-SA license.}

Send Feedback