IDIAP Head Pose dataset

The IDIAP Head Pose dataset [1] comes from 8 meeting sequences of 360x288 frame resolution, where two individuals were captured while discussing about various topics in a 4-person dialogue scenario. The total number of different subject captured is 15. They had their head orientations continuously annotated using a magnetic field location and orientation sensor tracker. The video repository has been employed for the CLEAR2007 head orientation estimation challenge, following the protocol described in [2]: 21152 75x75 samples were selected as training data and 23991 as testing data. Since the training samples are particularly unbiased on certain orientations, we flip them and then we randomly extract a subset of 5288 images, obtaining a balanced training pool. It represents a valuable benchmark since the annotations express the pan, tilt and roll angles of the head pose.


Dataset download: [IHDPHeadPose]

For questions regarding the data please contact diego[dot]tosato[at]univr[dot]it.


D. Tosato, M. Spera, M. Cristani, and V. Murino, Characterizing humans on
Riemannian manifolds, IEEE Transactions on Pattern Analysis and Machine
Intelligence (TPAMI), Accepted 2012. [pdf]

License Terms

This dataset is made available to the scientific community for non-commercial  research purposes such as academic research, teaching, scientific publications,  or personal experimentation. Permission is granted to use, copy, and distribute the data given that you agree:

1) That the dataset comes "AS IS", without express or implied warranty.  Although every effort has been made to ensure accuracy,  the University of Verona - VIPS lab,  as website host) does not accept any responsibility for errors or omissions.
2) That you include a reference to the above publication in any published  work that makes use of the dataset.
3) That if you have altered the content of the dataset or created derivative work,  prominent notices are made so that any recipients know that they are not receiving  the original data.
4) That you may not use or distribute the dataset or any derivative work for commercial purposes as, for example, licensing or selling the data, or using the data with a purpose to procure a commercial gain.
5) That this original license notice is retained with all copies or derivatives  of the dataset. That all rights not expressly granted to you are reserved by  University of Verona - VIPS lab.


[1] J. M. Odobez, “Idiap head pose database,”
[2] S. Ba and J. Odobez, “Evaluation of multiple cue head pose estimation algorithms  in natural environements,” in Proc.ICME. IEEE, 2005, pp. 1330–1333.