|
Entertainment Computing |
|
|
Autonomous World Simulator The user can particupate in the autonomous world, and experiment various story events through interactions with intelligent CG actors. |
|
|
Generating Digital Human from Video Building digital clones of movie actors are important for making CG animation and special effects in movies. We propose a new technique for learning and synthesizing actions of synthetic actors from movie sequences. First, we estimate actors body movement from a video sequence. Then we decompose the action sequence into the semantically meaningful segments such as walking and sitting. The new action can be synthesized using the state transition of action segments. We have appried the proposed method to the movie sequences, and generate new actions for CG characters. Keyeords: digital human, digital actor, model-based synthesis, human body tracking, human body fitting J. Hoshino:"Extracting
Personal Characteristics from Human Movement", Int. Conf. on Acaustics
and Signal Processing ( ICASSSP2001) , Conference Proceedings CD-ROM,
IMDSP-P6.1, 2001 [PDF]
|
|
|
Intelligent Storyboard Storyboard is a classical animation tool to help the creators to organize scenes. We are developpoing a new method for producing animation from storyboard. The 3D position and behavior of the characters are estimated from 2D views using the constraints optimization and example-based interpolation method. Junichi Hoshino, Yumi Hoshino:"Intelligent Storyboard for Prototyping Animation", IEEE Int. Conf. on Multimedia and Expo, ICME2001, Conference CD-ROM FAI.03 ,2001 [PDF] |
|
|
Interactive Virtual Fashion We are developping a Virtual fashion simulator by
merging video and CG. A person can wear any virtual cloth before the virtual
mirror. The synthesized virtual cloth and hair style can be matched with
the use's movement as if the user can weare them before the virtual mirror.
First, we track current 3D pose of human body by using the spatio-temporal
analysis. Then we generate CG objects and merge it with the human body
in video. (abstract)
(PDF). |
See-Through Representation of Video The operator can interactively control the transparency
of objects in video images, so that one can see behind the scenes of the
objects without occlusion. The see-through representation is useful for
many applications such as video surveillance, tele-operation of robots,
and virtual walkthrough.(abstract)
Junichi Hoshino:"See-Through
Representation of Video",2nd International Synposium on Mixed Reality,
ISMR2001, pp.155-156, 2001 [PDF]
Jun'ichi Hoshino, Tetsuya Uemura,
Isao Masuda: "Region-Based Reconstruction of an Indoor Scene Using an
Integration of Active and Passive Sensing Techniques", IEEE 3rd Int. Conf.
on Computer Vision, ICCV90, pp.568-572 [PDF]
| 2005-02-17 |
|
