Mainstage EDM Benchmark
收藏资源简介:
Mainstage EDM Benchmark数据集由约翰斯·霍普金斯大学、东南大学、国防科技大学和MBZUAI共同创建,旨在解决主流舞蹈音乐子流派分类的问题。该数据集包含超过1000首来自国际知名唱片公司的曲目,涵盖8个子流派,并采用连续软标签方法以捕捉跨子流派的特征。数据集的创建过程包括音频片段的收集、标注和特征提取,确保了数据的高质量和多样性。该数据集主要应用于音乐推荐、DJ曲目编排和互动多媒体等领域,旨在提升音乐信息检索的准确性和效率。
Mainstage EDM Benchmark dataset was co-created by Johns Hopkins University, Southeast University, National University of Defense Technology, and MBZUAI, aiming to tackle the classification task of mainstream dance music subgenres. This dataset contains over 1,000 tracks sourced from internationally renowned record labels, covering 8 subgenres, and adopts the continuous soft label methodology to capture cross-subgenre features. The dataset creation process includes audio clip collection, annotation and feature extraction, ensuring high data quality and diversity. This dataset is mainly applied in fields such as music recommendation, DJ track arrangement and interactive multimedia, with the goal of improving the accuracy and efficiency of music information retrieval.

- 1Benchmarking Sub-Genre Classification For Mainstage Dance Music约翰斯·霍普金斯大学, 东南大学, 国防科技大学, MBZUAI · 2024年



