DVS128 Gesture Dataset

Access the dataset that was used to build a real-time, gesture recognition system described in the CVPR 2017 paper titled “A Low Power, Fully Event-Based Gesture Recognition System."

 

Download dataset

Read the paper

About the dataset

This dataset was used to build the real-time, gesture recognition system described in the CVPR 2017 paper titled “A Low Power, Fully Event-Based Gesture Recognition System.” The data was recorded using a DVS128. The dataset contains 11 hand gestures from 29 subjects under 3 illumination conditions and is released under a Creative Commons Attribution 4.0 license.

A. Amir, B. Taba, D. Berg, T. Melano, J. McKinstry, C. Di Nolfo, T. Nayak, A. Andreopoulos, G. Garreau, M. Mendoza, J. Kusnitz, M. Debole, S. Esser, T. Delbruck, M. Flickner, and D. Modha, "A Low Power, Fully Event-Based Gesture Recognition System," 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Honolulu, HI, 2017.

Contents of the dataset

Each trial has two files: an data file (.aedat) containing the DVS128 events, and an annotation file (.csv) containing the start and stop times of each gesture.

Filenames identify the subject and illumination condition in each trial. For example, user10_fluorescent_led.aedat and user10_fluorescent_led_labels.csv contain gestures recorded from user10 under a combination of fluorescent and LED lighting.

Learn more about computer vision at IBM Research