Mona Lisa's mysterious smile created by using our method. The generated static painting looks similar to the original one when looking perpendicularly (a). However, when looking at the painting at different viewpoints, the expression of Mona Lisa's face may seem sadder or happier as shown in (b) and (c), and Mona Lisa's eyes may look around as shown in (d) and (e).
Abstract
This paper presents
a novel approach for creating self-moving objects using hollow-face
illusion. Given a clip of character animation, our approach generates a
static object. Looking at the object at different views, the similar
deformation can be observed. To accomplish this challenging mission, we
give qualitative and quantitative analysis of hollow-face illusion.
Methodology in computer vision and human perception are utilized to
design the algorithm. A static object is firstly generated to satisfy
the relative motion illusion constraints. The illusion is then
strengthened by back projecting the object to the 3D face space.
Considering both 'bottom-up' visual signal and 'top-down' knowledge, the
intended illusion can be generated. Experiments have shown the
effectiveness of our algorithm. For example, expression varying illusion
on an oil painting can be created by our method. The self-moving objects
can be used in applications such as design, entertainment, advertisement
and public safety.
What is illusion?
More about visual illusion: 105 Visual Phenomena & Optical Illusions
We were inspired from the Dragon Illusion (video).
http://www.grand-illusions.com/images/articles/opticalillusions/dragon_illusion/dragon_illusion.wmv
The generated static bunny using our algorithm, which seems to turn its head when the viewpoint changes.
3D printing of some experiment results (see live captured video)
We
would like to thank the anonymous reviewers for their constructive comments. We
thank Yu Shi and Linlin Xu for their help on the experiments. This work was
supported jointly by the National Natural Science Foundation of China
(61202284, 61170318, 61070071, 61222206).
title = {Mona Lisa alive: Create self-moving objects using hollow-face illusion},
author = {Jing Tong and Ligang Liu and Jin Zhou and Zhigeng Pan }
journal = {The Visual Computer (Proc. of Computer Graphics International)},
volume = {29},
number = {6-8},
pages = {535-544},
year = {2013}
}