Caricaturing

1
Caricaturing

• Ang Lee 2005/05/30

2
Objective

• Art creation auxiliary (bring new media to layman
  who cant draw/describe what in his mind)
• rather than have computer replace human artist.
• Input picture(s) of a subject, doesnt need
  expensive device (e.g. synchronous camera sets)
• Output Caricature of a subject
• (Mostly) automatic generate image
• Interactions happened when
• selecting among options (view angle, expressions,
  hair style)
• giving feed back (accept result/modify).
• Add/select other parts (body, background)
• but not manually adjusting details.

3
(Related) previous research Review

• PICASSO (3D)
• Developed by group in University of Chukyo, Japan
• Since early 1990s
• Modeling Expressive 3D Caricatures
• Ergun Akleman and Jon Reisch (T AM Univ)

SKIP Review and GO Directly to the method I
proposed ? (Click here to go to page 9)
4
PICASSO (3D)

• PICASSO is a big system there are several
  sub-systems of it
• 2D PICASSO
• metioned that they used side-view image
• but that paper is in Japanese only
• 3D PICASSO
• Interactive PICASSO
• Web PICASSO
• Emotional motion deformation
• Paper List
• paper (click) to an
• overview

5
3D PICASSO
Average Face
Compared with
Exaggerate
By human caricatursit
By 3D PICASSO
6
PICASSO paper lists (p1)

• From coarse to fine correspondence of 3-D facial
  images and its application to 3-D facial
  caricaturingKondo, T. Murakami, K. Koshimizu,
  H. ( University of Chukyo )3-D Digital Imaging
  and Modeling, 1997. Proceedings., International
  Conference on Recent Advances in12-15 May 1997
  Page(s)283 - 288
•  
• On autonomous control method for facial
  caricature generation based on a model of visual
  illusion-experimental modeling of visual
  illusionMurakami, K. Takai, M. Koshimizu,
  H.Advanced Intelligent Mechatronics '97.,
  IEEE/ASME International Conference on16-20 June
  1997 Page(s)29
•  
• Facial caricaturing with motion caricaturing in
  PICASSO systemTominaga, M. Fukuoka, S.
  Murakami, K. Koshimizu, H.Advanced Intelligent
  Mechatronics '97., IEEE/ASME International
  Conference on16-20 June 1997 Page(s)30
•  
• Facial caricaturing system PICASSO with emotional
  motion deformationTominaga, M. Hayashi, J.-I.
  Murakami, K. Koshimizu, H.Knowledge-Based
  Intelligent Electronic Systems, 1998. Proceedings
  KES '98. 1998 Second International Conference
  onVolume 2,  21-23 April 1998 Page(s)205 - 214
  vol.2
•  
•  New Kansei machine vision application. A
  prospect for human sensory factors in machine
  visionKoshimizu, H. Nagata, N.Knowledge-Based
  Intelligent Electronic Systems, 1998. Proceedings
  KES '98. 1998 Second International Conference
  onVolume 2,  21-23 April 1998 Page(s)185 - 192
  vol.2
•  
• On the detection of feature points of 3D facial
  image and its application to 3D facial
  caricatureFujiwara, T. Nishihara, T. Tominaga,
  M. Kato, K. Murakami, K. Koshirnizu, H.3-D
  Digital Imaging and Modeling, 1999. Proceedings.
  Second International Conference on4-8 Oct. 1999
  Page(s)490 496
•  
• On KANSEI facial image processing for
  computerized facial caricaturing system
  PICASSOKoshimizu, H. Tominaga, M. Fujiwara,
  T. Murakami, K. (University of Chukyo, Aichi
  Perfectural University)Systems, Man, and
  Cybernetics, 1999. IEEE SMC '99 Conference
  Proceedings. 1999 IEEE International Conference
  onVolume 6,  12-15 Oct. 1999 Page(s)294 - 299
  vol.6
•  
• Dynamic facial caricaturing system based on the
  gaze direction of galleryMurakami, K. Tominaga,
  M. Koshimizu, H.Automatic Face and Gesture
  Recognition, 2000. Proceedings. Fourth IEEE
  International Conference on28-30 March 2000
  Page(s)136 141
• http//ieeexplore.ieee.org/Xplore/guesthome.jsp

7
PICASSO paper lists (p2)

• An interactive facial caricaturing system based
  on the gaze direction of galleryMurakami, K.
  Tominaga, M. Koshimizu, H.Pattern Recognition,
  2000. Proceedings. 15th International Conference
  onVolume 4,  3-7 Sept. 2000 Page(s)710 - 713
  vol.4
•  
• KANSEI facial caricaturing based on the
  eye-camera interface from galleryTominaga, M.
  Asao, T. Murakami, K. Koshimizu, H.Industrial
  Electronics Society, 2000. IECON 2000. 26th
  Annual Confjerence of the IEEEVolume 1,  22-28
  Oct. 2000 Page(s)307 - 312 vol.1
•  
• 3D modeling system of human face and full 3D
  facial caricaturingFujiwara, T. Koshimizu, H.
  Fujimura, K. Kihara, H. Noguchi, Y. Ishikawa,
  N.3-D Digital Imaging and Modeling, 2001.
  Proceedings. Third International Conference on28
  May-1 June 2001 Page(s)385 - 392
•  
• 3D modeling system of human face and full 3D
  facial caricaturingFujiwara, T. Koshimizu, H.
  Fujimura, K. Fujita, G. Noguchi, Y. Ishikawa,
  N.Virtual Systems and Multimedia, 2001.
  Proceedings. Seventh International Conference
  on25-27 Oct. 2001 Page(s)625 633
•  
• A method for 3D face modeling and caricatured
  figure generationFujiwara, T. Koshimizu, H.
  Fujimura, K. Fujita, G. Noguchi, Y. Ishikawa,
  N.Multimedia and Expo, 2002. ICME '02.
  Proceedings. 2002 IEEE International Conference
  onVolume 2,  26-29 Aug. 2002 Page(s)137 - 140
  vol.2
•  
• Hierarchical face tracking by using PTZ
  cameraFunahasahi, T. Tominaga, M. Fujiwara,
  T. Koshimizu, H.Automatic Face and Gesture
  Recognition, 2004. Proceedings. Sixth IEEE
  International Conference on17-19 May 2004
  Page(s)427 432
•  http//ieeexplore.ieee.org/Xplore/guesthome.jsp

8
Modeling Expressive 3D CaricaturesSiggraph
Sketch 2004

• (Authors Slides for Siggraph Talk (ppt))
• Description(Sketch (pdf) )

abstract caricatures Created using disconnected
piece by HUMAN
C Maya API
Result
9
motivation

• Most current methods are restricted to 2D image
  processing. Hence, the issues of perspective is
  not considered.
• Though in final work perspective is often not
  preserved. But when dealing with how to distorted
  or warp image should take perspective into
  concern.
• For example, a mans lower part is shorter (nose,
  mouth, and chin). When dealing with How to
  exaggerate and shorten each subparts further,
  perspective issue need to be considered.

Linearly distort
10
motivation

• Think Three-dimensionally If theres a deep
  indenture under the cheekbones, make sure your
  drawing show it
• Lenn Redman ltHow to draw caricaturesgt

11
Why need side view?

• Currently only front images are input
• Can you tell his chin protrudes ?

12
Front View?

• Moreover, as a result, current methods, the
  results are restricted to front view

Which one is better?
13
(No Transcript)
14
3D model without texture?

• Caricature of Marilyn Monroe without a mole?
• Caricature of Saddam Hussein without a mustache?
• Einstein without wrinkles on the forehead?

15
Angs method Flow diagram
Average Head 3D model
Modify the model
Similar work by artist
Proximate 3D model of the subject
Analysis
Texture
Texture
Texture Skin on the model
Exaggerate automatically
Change facial expression if needed
Rotate, Zoom In out (perspective), scalemodify
view angle
Expected results
Add accessory hair, glasses if needed
3D to 2D
Image processing to different styles (oil
painting, cartoon, sketch)
16
Issues to be dealt with

• Interaction design. (Which part to human which
  parts to machine?)
• Get an average head 3D model.
• How to modify the average head 3D model base on
  front and side view images?
• How to put texture onto the model?
• How to exaggerate?
• How to add accessory parts? (hair style, glasses)
• Detect and change facial expression (emotions).