The LIRIS human activities dataset - download

You need to cite the following journal paper in all publications including results for which you used the LIRIS dataset:.

C Wolf, J. Mille, E. Lombardi, O. Celiktutan, M. Jiu, E. Dogan, G. Eren, M. Baccouche, E. Dellandrea, C.-E. Bichot, C. Garcia, B. Sankur, Evaluation of video activity localizations integrating quality and quantity measurements, In Computer Vision and Image Understanding (127):14-30, 2014.

Description of fileformats.

Subset D1 (Mobile robot + Kinect)

• Training data : grayscale frames + depth frames (2.8 GB) [TAR]
• Training data (OPTIONAL!) : pseudocolor visualization of depth data (1.3 GB, not required) [TAR]
• Training data : XML annotations (bounding boxes and action classes). The coordinates of the bounding boxes are for the grayscale frames, not the depth frames. [TGZ]
• Test data : grayscale frames + depth frames (1.7 GB) [TAR]
• Test data : XML annotations (bounding boxes and action classes). The coordinates of the bounding boxes are for the grayscale frames, not the depth frames. [TGZ]

ATTENTION!! Due to a glitch in the sampling and streaming software, there is a temporal offset in the gray and depth videos. Frame X in the gray video corresponds to frame X-1 in the depth video! We apologize for this.

Subset D2 (Sony camcorder)

• Training data file 1/3 (3.6 GB) [TAR]
• Training data file 2/3 (1.9 GB) [TAR]
• Training data file 3/3 (3.7 GB) [TAR]
• Training data : XML annotations (bounding boxes and action classes) [TGZ]
• Test data file 1/2 (2.9 GB) [TAR]
• Test data file 2/2 (1.5 GB) [TAR]
• Test data : XML annotations (bounding boxes and action classes) [TGZ]

Actreader : read and browse annotated videos

• Documentation
• Download script for all systems [TGZ]

Actanno : edit and create annotations

• Documentation (installation and usage)
• Download for Linux / MacOS [TGZ]
• Download for Windows [ZIP]

(A new version of actanno has been developed by Joris Guerry at ONERA. It is available on github, but uses a slightly different output XML format.)

Kinect calibration and correspondence software

We also provide calibration details of the Kinect module used during acquisition and software calculating the correspondence between the depth images and the grayscale images.

• Documentation (including parameter values)
• Download of correspondence software [TGZ]

Harleval - evaluation tool

• Documentation (usage manual)
• Documentation of the metric
• Download [ZIP]

Organization, contact

The dataset was collected and produced by members of the LIRIS Laboratory, CNRS, France
Christian Wolf, Julien Mille, Eric Lombardi, Bülent Sankur (BUSIM, Bogazici University, Turkey), Emmanuel Dellandréa, Christophe Garcia, Charles-Edmond Bichot, Mingyuan Jiu, Oya Celiktutan, Moez Baccouche

Send questions to christian.wolf (at) liris.cnrs.fr

You need to cite the following journal paper in all publications including results for which you used the LIRIS dataset:.

C Wolf, J. Mille, E. Lombardi, O. Celiktutan, M. Jiu, E. Dogan, G. Eren, M. Baccouche, E. Dellandrea, C.-E. Bichot, C. Garcia, B. Sankur, Evaluation of video activity localizations integrating quality and quantity measurements, In Computer Vision and Image Understanding (127):14-30, 2014.