Pablo Picasso according to wikipedia ... fauvism complimentary - - PowerPoint PPT Presentation
Pablo Picasso according to wikipedia ... fauvism complimentary - - PowerPoint PPT Presentation
The Fauvist Works of Pablo Picasso according to wikipedia ... fauvism complimentary colours bold simplified strong heavy abstracted textured brush strokes wide visible paintings of Marie-Thrse Walter by Pablo Picasso circa 1932
fauvism
strong
brush strokes
complimentary
colours
bold wide heavy visible textured
simplified abstracted
according to wikipedia ...
the inspiration
paintings of Marie-Thérèse Walter by Pablo Picasso circa 1932
combination of f auvist and cubist elements
previous work
“Cubist Style Rendering from Photographs”
J.P. Collomosse and P.M. Hall
- general salient f
eatures identified
- planar f
eatures distorted using vector fields
- multiple perspectives achieved by use of multiple images
- assemble planes using image registration
- painterly render
characterization design future work
contents
characterization
shape
flowing curves bulbous exaggeration pinching thick black outlines simplification circles
characterization
colour
bright complimentary reversed surreal solid note - the artists colour wheel is red, yellow, blue ...
characterization
style
women patterns f ew gradients not exceptionaly painterly rough canvas texture dry brush additional lines
design
directly related to characterization three major components to consider shape
detect distort
colour
discover choose
style
strokes texture
components are seperable!
design
pre-process segment
edge discovery
set of curves
shape
distortion find circles local pinch
rot/scale/trans
extra lines eg - segmentation input image
design
to find circles ...
find center
C x=∑ xi Nx ,C y=∑ yi Ny
find radius
r=∑∣P−C∣
2N
if f
- r all points P
valid++
∣r −k∣≤∣P−C∣≤∣rk∣
if circle! not circle!
validallowed valid≥allowed
point set
eg
design
ini tial point set polygonal shape
center point
result
approximating circle a=50% k=15 *actual output
point set
compute θ
=arccos PQ⋅ PR
θ P Q R compute d f
- r P, Q, & R
d=∣ ∣C−P∣ ∣
2d=C x−Px
2Cy−P y 2if move P
P0=Pt C−P t= 1 2
split edge
E={e0,e1,... ,en} E0={e0,e1,... ,ei ,P0} E1={P0,ei1,...,en}
continue while more points
dP≤dR∧d P≤d Q∧≤min
set of edges no more points
design
local pinch ...
- riginal curve
point P shif ted pinched curve set P P P C C C Q R Q Q R R θ
design
eg
local pinching ...
*actual output
design
colour
enumerate colours group by major hue
RG∧RB GB∧GR BG∧BR ∣B−G∣≤30∧minB,GR ∣R−G∣≤30∧minR,GB ∣R−B∣≤30∧minR,BG
remove grays combine colours
∣ ∣C1−C2∣ ∣2≤k
BC1BC2 delete BC2
input paintings possible palette outputs ...
design
colour
draw cruves re-segment
- curves associated wi
th segments
- controls segment colour af
ter distortions
paint pixels paint patterns painterly render
paint black edges texturize
- paint flat colours
design
- pick regions of shadow and patternize
- alpha blending using
canvas texture
style
future work
- segmentation algori
thms and f eature recognition to allow characteristic distortion of specific f eatures such as eyes
- 3D models instead of photographs could allow f
- r
f eature recogni tion, shadow detection and perspective distorions
- extra distortions such as addi
tion of extra lines, shape cutting, intelligent replication, ...
WYSIWY G NPR: Drawing Strokes Directly on 3D Models
Kalnins, Markosian, Meier, Kowalski, Lee, Davidson, Webb, and Hughes
Cubist Style Rendering from Photographs
Collomosse and Hall