DART: Articulated Hand Model with Diverse Accessories and Rich Textures
Daiheng GaoYuliang XiuKailin LiLixin YangFeng WangPeng ZhangBang ZhangCewu LuPing Tan
Daiheng GaoYuliang XiuKailin Li
...+5
Ping Tan
Oct 2022
0被引用
12笔记
读论文，拿好礼活动火爆进行中，iPad、蓝牙耳机、拍立得、键盘鼠标套装等你来拿！
摘要原文
Hand, the bearer of human productivity and intelligence, is receiving muchattention due to the recent fever of digital twins. Among different handmorphable models, MANO has been widely used in vision and graphics community.However, MANO disregards textures and accessories, which largely limits itspower to synthesize photorealistic hand data. In this paper, we extend MANOwith Diverse Accessories and Rich Textures, namely DART. DART is composed of 50daily 3D accessories which varies in appearance and shape, and 325 hand-crafted2D texture maps covers different kinds of blemishes or make-ups. Unity GUI isalso provided to generate synthetic hand data with user-defined settings, e.g.,pose, camera, background, lighting, textures, and accessories. Finally, werelease DARTset, which contains large-scale (800K), high-fidelity synthetichand images, paired with perfect-aligned 3D labels. Experiments demonstrate itssuperiority in diversity. As a complement to existing hand datasets, DARTsetboosts the generalization in both hand pose estimation and mesh recovery tasks.Raw ingredients (textures, accessories), Unity GUI, source code and DARTset arepublicly available at dart2022.github.io
