期刊
JOURNAL OF COMPUTING AND INFORMATION SCIENCE IN ENGINEERING
卷 22, 期 5, 页码 -出版社
ASME
DOI: 10.1115/1.4054001
关键词
artificial intelligence; computer aided design; virtual prototyping
资金
- HKSAR Innovation and Technology Commission (ITC) Innovation and Technology Fund [ITT/032/18GP]
- Natural Science Foundation of China [61976040, 61702079, 61762064]
- Science and Technology Development Fund of Macau SAR [SKL-IOTSC-2018-2020, 0018/2019/AKP, 0008/2019/AGJ]
In this article, a supervised learning approach based on CNN is investigated to solve the problem of 3D human modeling from two orthogonal silhouettes, proposing a new network structure to enhance model accuracy.
In this article, we revisit the problem of 3D human modeling from two orthogonal silhouettes of individuals (i.e., front and side views). Different from our previous work (Wang et al. (2003, Virtual Human Modeling From Photographs for Garment Industry, Comput. Aided Des., 35, pp. 577-589).), a supervised learning approach based on the convolutional neural network (CNN) is investigated to solve the problem by establishing a mapping function that can effectively extract features from two silhouettes and fuse them into coefficients in the shape space of human bodies. A new CNN structure is proposed in our work to extract not only the discriminative features of front and side views but also their mixed features for the mapping function. 3D human models with high accuracy are synthesized from coefficients generated by the mapping function. Existing CNN approaches for 3D human modeling usually learn a large number of parameters (from 8.5 M to 355.4 M) from two binary images. Differently, we investigate a new network architecture and conduct the samples on silhouettes as the input. As a consequence, more accurate models can be generated by our network with only 2.4 M coefficients. The training of our network is conducted on samples obtained by augmenting a publicly accessible dataset. Learning transfer by using datasets with a smaller number of scanned models is applied to our network to enable the function of generating results with gender-oriented (or geographical) patterns.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据