Dance with Melody Dataset
收藏资源简介:
该数据集旨在建立舞蹈动作与音乐之间的关系,包含四种舞蹈类型:恰恰、探戈、伦巴和华尔兹。每个目录中的`audio.mp3`提供舞蹈的音频文件,`skeletons.json`描述舞蹈的骨骼点,`config.json`给出舞蹈序列匹配歌曲部分的开始帧和结束帧。数据集的帧率为25帧每秒。
This dataset is designed to establish the relationship between dance movements and music, encompassing four dance genres: Cha-Cha, Tango, Rumba, and Waltz. Each directory contains an `audio.mp3` file providing the audio for the dance, a `skeletons.json` file describing the skeletal points of the dance, and a `config.json` file indicating the start and end frames of the dance sequence that match the song sections. The dataset operates at a frame rate of 25 frames per second.
数据集概述
数据集介绍
本数据集旨在构建舞蹈动作与音乐之间的关系,包含四种舞蹈类型:恰恰、探戈、伦巴和华尔兹。每个目录中,audio.mp3 包含舞蹈的音频文件,skeletons.json 描述舞蹈的骨骼点,config.json 提供舞蹈序列的起始帧和结束帧,数据集的帧率为每秒25帧。
数据集内容
- 舞蹈类型:恰恰、探戈、伦巴、华尔兹
- 文件格式:
audio.mp3:音频文件skeletons.json:骨骼点描述config.json:起始帧和结束帧配置
- 帧率:25 FPS
相关论文
-
论文1:
- 标题:Dance with Melody: An LSTM-autoencoder Approach to Music-oriented Dance Synthesis
- 作者:Tang, Taoran and Jia, Jia and Mao, Hanyang
- 会议:2018 ACM Multimedia Conference on Multimedia Conference
- 页码:1598-1606
- 年份:2018
-
论文2:
- 标题:AniDance: Real-Time Dance Motion Synthesize to the Song
- 作者:Tang, Taoran and Mao, Hanyang and Jia, Jia
- 会议:2018 ACM Multimedia Conference on Multimedia Conference
- 页码:1237-1239
- 年份:2018




