SIZER
收藏资源简介:
SIZER数据集是由马克斯普朗克信息学研究所创建的,包含约2000个不同尺寸的休闲服装扫描数据,涵盖100个不同体型的穿着者。该数据集不仅包括扫描数据,还提供了对SMPL模型的注册、服装部分的分割、服装类别和尺寸标签。创建过程涉及对3D扫描的分割和体型估计,以及对服装的跨数据集注册。SIZER数据集主要应用于虚拟试衣、角色动画、生成模型学习等领域,旨在解决3D服装建模中尺寸和体型的敏感性问题。
The SIZER dataset was developed by the Max Planck Institute for Informatics. It comprises approximately 2,000 scanned casual clothing samples of varying sizes, covering 100 wearers with distinct body shapes. In addition to the raw scan data, this dataset also provides SMPL model registration, clothing part segmentation, as well as clothing category and size annotations. The dataset creation process involves 3D scan segmentation and body shape estimation, alongside cross-dataset registration for clothing items. The SIZER dataset is primarily applied in fields such as virtual try-on, character animation, generative model learning and other relevant domains, aiming to address the sensitivity issues of size and body shape in 3D clothing modeling.

- 1SIZER: A Dataset and Model for Parsing 3D Clothing and Learning Size Sensitive 3D Clothing马克斯普朗克信息学研究所 · 2020年



