- 最后登录
- 2018-6-29
- 注册时间
- 2011-7-1
- 阅读权限
- 20
- 积分
- 359
- 纳金币
- 335582
- 精华
- 0
|
Compression and Direct Manipulation of Complex Blendshape Models
Jaewoo Seo Geoffrey Irvingy J. P. Lewisz Junyong Nohx
KAIST Weta Digital Weta Digital KAIST
Abstract
We present a method to compress complex blendshape models and
thereby enable interactive, hardware-accelerated animation of these
models. Facial blendshape models in production are typically large
in terms of both the resolution of the model and the number of tar-
get shapes. They are represented by a single huge blendshape ma-
trix, whose size presents a storage burden and prevents real-time
processing. To address this problem, we present a new matrix com-
pression scheme based on a hierarchically semi-separable (HSS)
representation with matrix block reordering. The compressed data
are also suitable for parallel processing. An efficient GPU imple-
mentation provides very fast feedback of the resulting animation.
Compared with the original data, our technique leads to a huge im-
provement in both storage and processing efficiency without incur-
ring any visual artifacts. As an application, we introduce an ex-
tended version of the direct manipulation method to control a large
number of facial blendshapes efficiently and intuitively.
CR Categories: I.3.7 [Computer Graphics]: Three-Dimensional
Graphics and Realism—Animation;
Keywords: blendshape, compression, direct manipulation
1 Introduction
Blendshapes are a predominant technique for the creation of real-
istic and semi-realistic facial animation. The strength of the blend-
shape method lies in its simple and interpretable parameterization:
the artist defines an ‘expression space’ with shapes such as raise-
inner-brow-right, and then models each shape exactly as desired.
Blendshape deformers are available in most commercial graphics
packages such as Autodesk maya and 3ds max, and virtual charac-
ters created with this technique abound in recent popular films.
全文请下载附件:
|
|