MMFS
收藏资源简介:
MMFS数据集是由大连理工大学计算机科学与技术学院创建,专注于花样滑冰的细粒度动作分析。该数据集包含11671个视频片段,涵盖256种动作类别,并提供了RGB和骨骼数据等多种模态信息。数据集的构建过程严格,通过专家和官方文档的联合标注确保了数据质量。MMFS数据集的应用领域主要集中在动作识别和动作质量评估,旨在解决现有数据集在动作细粒度分析上的不足,推动相关技术的发展。
The MMFS dataset was created by the School of Computer Science and Technology, Dalian University of Technology, and focuses on fine-grained action analysis of figure skating. It contains 11,671 video clips covering 256 action categories, and provides multi-modal information such as RGB data and skeleton data. The dataset is constructed rigorously, with its data quality ensured through joint annotation by experts and official documents. The main application fields of the MMFS dataset are action recognition and action quality assessment, aiming to address the shortcomings of existing datasets in fine-grained action analysis and promote the development of related technologies.

- 1Fine-grained Action Analysis: A Multi-modality and Multi-task Dataset of Figure Skating大连理工大学计算机科学与技术学院 · 2024年



