DrapeMoCapBench (DMCB)

DMCB是一个专为评估光学标记式和无标记式运动捕捉(MoCap)方法性能而设计的基准数据集。该数据集利用大规模真实世界MoCap数据，通过3D物理模拟，涵盖了从紧身到极宽松衣物的六种垂坠度、三种运动级别及六种体型-性别组合，以评估和比较不同场景下的MoCap方法。DMCB旨在帮助智能穿戴研究者选择最适合其特定应用的运动捕捉方法，平衡衣物设计、运动类型、成本和时间开销及精度等因素。DMCB is a benchmark dataset specifically developed to assess the performance of optical marker-based and markerless motion capture (MoCap) methods. Utilizing large-scale real-world MoCap data and incorporating 3D physical simulation, this dataset covers six clothing drape levels ranging from tight to extremely loose, three motion intensity levels, and six body type-gender combinations, to facilitate the evaluation and comparison of MoCap methods across diverse application scenarios. DMCB is designed to help wearable technology researchers select the optimal motion capture method for their specific use cases, balancing factors including clothing design, motion type, cost, time overhead and capture accuracy.