POSTURES-DVS Database

Teresa Serrano-Gotarredona

and

Bernabé Linares-Barranco

Instituto de Microelectrónica de Sevilla, IMSE-CNM (CSIC & Univ. of Sevilla), Sevilla, SPAIN.

         

                                            Bending                                                       Hand1                                                Hand2

                    

                                            Squat                                                        Swing                                                 Stand

We present a dataset of human postures recorded with a DVS sensor:

POSTURES-DVS

The POSTURES-DVS database consists of DVS recordings of two individuals adopting 6 different body postures named as: `bending’,`hand1’,`hand2’,`squat’,`swing’ and `stand’.

This dataset has been used in

S. Chen, P. Akselrod, B. Zhao, J. A. Pérez-Carrasco, B. Linares-Barranco and E. Culurciello, "Efficient feedforward categorization of objects and human postures with address-event image sensors," IEEE Trans. on Pattern Analysis and Machine Intelligence, vol. 34, No. 2, pp. 302-314, Feb. 2012. (ieeexplore)

The recordings have been obtained using the sensitive DVS developed in IMSE

T. Serrano-Gotarredona and B. Linares-Barranco, "A 128x128 1.5% Contrast Sensitivity 0.9% FPN 3us Latency 4mW Asynchronous Frame-Free Dynamic Vision Sensor Using Transimpedance Amplifiers," IEEE J. Solid-State Circuits, vol.48, No. 3, pp. 827-838, March 2013 (available from:

 http://ieeexplore.ieee.org/xpl/articleDetails.jsp?arnumber=6407468).

Recorded files can be displayed with jAER (available from: http://sourceforge.net/p/jaer/wiki/Home). MatlabTM scripts are provided to transfer files from/to jAER format to matlab.

 

This database can be downloaded free of charge, provided it is used for non-commercial purposes and the original source is credited in publications and reports.

To download it go to http://imse-cnm.csic.es/caviar/postures_dvs.