Title

News (November 6th, 2019)

We just started distributing the AIST Dance Video Database (AIST Dance DB)! This first release contains all 13,940 original videos ("raw" videos) with 60 original musical pieces. Since these raw videos were directly taken from video cameras, each of them consists of a pre-roll, main dancing part, and post-roll, and is accompanied by noisy raw audio recordings. We are preparing the second release that includes "refined" videos. Each refined video is generated by trimming away the pre-roll and post-roll of a raw video and replacing the raw audio with its noiseless original musical piece. The second release will be available by the end of November 2019. As a preview, you can see some refined videos at a "compact subset" on Database Download.

Overview

AIST Dance Video Database (AIST Dance DB) is a shared database containing original street dance videos with copyright-cleared dance music. This is the first large-scale shared database focusing on street dances to promote academic research regarding Dance Information Processing. The AIST Dance DB will foster a variety of new tasks such as

  • Dance-motion genre classification
  • Dancer identification
  • Dance-technique estimation
This database features:

Multiple genres

(ten major dance genres)
a

Multiple dancers

(solo and group dancing by 40 professional dancers)
a

Multiple cameras

(at most nine video cameras surrounding a dancer)
a

How to Get Started

Please read through the Terms of Use before you start using the database. The Terms of Use contains terms and conditions for using the database as well as disclaimers. The structure of the database is presented in detail at Database Structure.

There are several ways to download the database depending on your needs. You can download the full database (515 Gigabytes) from Getting the database. To download a subset of the database, you can choose the Filtered Search option. The musical pieces (original dance music) used in the database are also available at Get Only Music Data.

Along with the database, we have proposed a new task named dance-motion genre classification as a benchmark task. The subset data for this task is available at Tasks.

It is necessary to cite the following publication when you use AIST Dance DB.

Citation

Shuhei Tsuchida, Satoru Fukayama, Masahiro Hamasaki and Masataka Goto. AIST Dance Video Database: Multi-genre, Multi-dancer, and Multi-camera Database for Dance Information Processing. In Proceedings of the 20th International Society for Music Information Retrieval Conference (ISMIR 2019), 2019.

http://archives.ismir.net/ismir2019/paper/000060.pdf

@inproceedings{aist-dance-db,
           author = {Shuhei Tsuchida and Satoru Fukayama and Masahiro Hamasaki and Masataka Goto}, 
           title = {AIST Dance Video Database: Multi-genre, Multi-dancer, and Multi-camera Database for Dance Information Processing}, 
           booktitle = {Proceedings of the 20th International Society for Music Information Retrieval Conference, {ISMIR} 2019},
           address = {Delft, Netherlands}, 
           year = 2019, 
           month = nov }

Acknowledgments

AIST Dance DB was supported in part by JST ACCEL Grant Number JPMJAC1602, Japan. ANOMALY INC. was in charge of music production and dance performance recording. Takahiro Inoue manages the database server. Fabrizio Pedersoli and Hirohito Tanaka helped to prepare the refined videos in the database. The logo was designed by Misato Nakamura.