IIITA -ROBITA Indian Sign Language Gesture Database


Introduction

Recognizing and Interpreting Indian Sign Language Gesture has been becoming an active research area in the field of Human Robot Interaction. The Indian Institute of Information Technology Allahabad (IIITA)- ROBITA will provide IIITA-ROBITA ISL Gesture Database to gesture recognition researchers in order to promote further development and studies in this research. The IIITA-ROBITA has copyright in the collection of ISL gesture videos and image data and serves as a distributor of the IIITA-ROBITA ISL gesture database.

The data have been created since July 2009 at the Robotics and AI Lab, IIIT-Allahabad. The data set consists of 23 different gestures captured at 30 fps, 320 by 240 pixels with Sony Handycam. We have maintained constant background with various light illumination conditions. Detailed descriptions are found in the following papers.

1) A. Nandy, S. Mondal, J.S.Prasad, P. Chakraborty and G. C. Nandi - Recognizing & Interpreting Indian Sign Language Gesture for Human Robot Interaction - In the proceeding of ICCCT-10, IEEE Xplore Digital Library, September 2010, pp. 712-717.

2) A. Nandy, S. Mondal, J.S.Prasad, P. Chakraborty and G. C. Nandi - Recognition of Isolated Indian Sign Language gesture in Real Time - In the book of (Information Processing and Management) Springer LNCS-CCIS, March 2010, Vol. 70, pp. 102-107.

Datasets Description

The ISL gesture dataset is composed of sequences of RGB frames for 23 Isolated ISL gestures. Samples of still images are shown below (i.e. above, across, afraid, etc).


Above Gesture


Across Gesture


Afraid Gesture


Advance Gesture


How to get the dataset?

The entire dataset would be downloaded as a password protected zip file and the necessary password will be issued on subject to subject basis. In order to obtain the password, one must send the release agreement signed by an official representative of your institution (for instance, your supervisor if you are a student) to the given email id mentioned on the release agreement.


Note -

Please use separate unzip software (i.e. 7-Zip) if necessary to extract all the gesture files