AUTOMATIC MESH ANIMATION PREVIEW
Yi Li, Qiaodong Cui, Fei Dou, Lin Zhang, Zhong Zhou*
State Key Laboratory of Virtual Reality Technology and Systems, Beihang University
Beijing, China
*e-mail: zz@vrlab.buaa.edu.cn
ABSTRACT
With growing number of high quality 3D models published
online, the technique of generating efficient and economical
3D model previews has raised increasing concerns. Although
several previous work has been done on the preview of stat-
ic mesh, that of animated mesh is different and more complex
due to the difficulty in describing the animation. In this paper,
we present a novel method of automatic preview generation
for 3D mesh animation. A new measure named inter-frame
surface saliency, which evaluates both inter-frame motions
and surface saliency in each frame, is introduced. Given an
animated mesh, an energy function combining this measure
and camera smoothness is constructed for the representative
viewpoints selection in key frames, and then an optimal cam-
era path is generated. Finally, a brief but informative preview
could be created by moving the camera along this path with
frame rate control.
Index Terms— mesh animation preview, inter-frame sur-
face saliency, camera path generation
1. INTRODUCTION
With the development of 3D modeling and animation tech-
niques, as well as the computer hardware, the amount of high
quality 3D static/animated models is increasing at a relatively
high rate. Today there have been more and more online data
publishers for 3D models, such as TurboSquid, The3DStudio
and Berkeley MHAD, which allow users to download mod-
els or animation datas for academic or commercial purposes.
An efficient and economical 3D static/animated models pre-
view technique is usually needed to give the user a general
understanding before download, due to the growing data size
of 3D models. Many works have been done on preview of
3D static model, however that of animated model is different
and more complex due to the difficulty of creating a brief but
informative description of the animation.
This work is supported by the National 863 Program of China under
Grant No.2012AA011801, the Natural Science Foundation of China under
Grant No.61170188, and the Specialized Research Fund for the Doctoral Pro-
gram of Higher Education of China under Grant No.20121102130004.
Similar to the preview of static model, there are main-
ly two kinds of methods to preview models, image-based and
video-based. The idea of image-based method is using a set of
images taken from different representative viewpoints around
a 3D model to help users understand it [1, 2, 3, 4]. The video-
based method is to preview the model through a video gener-
ated from a selected camera path [5, 6, 7, 8, 9]. Comparing
to the image-based method, the video-based one gives a bet-
ter view of the continuous dynamic changes of the animated
models, resulting a better understanding of the animation.
To the best of our knowledge, most of the image-based
and video-based methods work on static models, few work-
s have been done on mesh animation preview before. Han
et al. [5] proposed a method to generate automatic preview
video of mesh sequences by adopting Dijkstra’s algorithm
which considering both of information quantity and camer-
a travel distance. However, Han et al.’s method considered
only the information quality from the static models’ point of
view without taking the significance of motion into accoun-
t. Moreover, this method is actually an image-based strategy
that only selects a fixed viewpoint for each frame and there-
fore cannot generate a real smooth preview video.
In this paper, we present a novel method of automatic
preview video generation for 3D mesh animations. A new
measure, namely inter-frame surface saliency, which contain-
s both of motion information between frames and surface
saliency in each frame is introduced. To generate a brief but
informative description of the animation sequences, some key
frames of the animated mesh are firstly selected. An ener-
gy function of both inter-frame surface saliency and camera
smoothness is constructed for computing the representative
viewpoints, and an optimal camera path is then generated. Fi-
nally, a brief preview video is created by moving the camera
along this path with frame rate control.
The rest of this paper is organized as follows. In Sec-
tion 2, we present the preprocessing steps of viewpoint se-
lection called motion analysis and the definition of our new
measure. The algorithm of optimal camera path generation is
presented in Section 3. The evaluation of experiment results
on several models are presented in Section 4 and followed by
a conclusion and expectation in Section 5.