simple graphic palette
palette
Switches to a very simple 2D palette mode which cannot replace the wealth of a
programmable palette:
multiple brushes, automatic, dynamic, possibly connected to sensors via neural networks,
etc..
Options
back(r,g,b): background color (black default).
coe(cx,cy): when displayed the sizes
of the image can be choosen on 2 scales in [cx,DX/cx] and [cy,DY/ccy]
(with DX,DY = sizes of thz screen).
directory(): directory (ima default).
name("*.eee": eee is the
extension of the images (bmp by default).
secx(x1,x2,...)secy(y1,y2,...): when displayed the sizes
of the image can be choosen on 2 scales nx=x1,x2,... and ny=y1,y2,....
By default (64,128,192,256,320,384,448,512,576,640,704,768,832,896,960,1024,1088).
Menu
The following menu is displayed by pressing the right button of the mouse:
help
displays the help.
val
validate the current operation.
read image
loads image files:
click on name or type a name: loads file name.
click on num: copy this file in image num, the current image
number is incremented.
click on reset images: resets all images to their initial values.
define brush
click on num_brush.
type
col
transp
radius
rotz
diat
alea
rand
type1:
disk:: circular.
rectangle: rectangular.
win: the image generated by win
is associated to the brush.
(must be window).
line: the image generated by lin
is associated to the brush.
(must be line).
type2:
nul: no type2
smooth: smooth:
click on smooth: size of the convolution matrix.
click on num_ima: number of the image associated to the brush.
texture: texture:
click on num_ima: number of the image associated to the brush.
network_1: luminance neural network
(must be define network lum).
network_3: color neural networks
(must be define network color).
click on image_network:
if num > 0: the network takes his pixel informations
in the image num (image brush(id)=num: avoiding brush overlap).
if -1: the network takes his pixel informations
in image num_ima itself (producing brush overlap).
bru_col:
1: r1,g1,b1 = color.
2: r1,g1,b1, r2,g2,b2 = color at center and on the edge.
bru_transp:
1: t1 = transparency.
2: t1,t2 = transparency at center and on the edge.
rx,ry: radius.
alea: reproductible random.
rand: no reproductible random.
rotz: z rotation.
dist: minimum capture distance.
network lum
Defines a network, change the scales:
nx,ny: x,y sizes.
hide: number of hidden layers (increase if no convergence).
nb: number of learning couples.
rand: randomly generate motif and law number num.
The motifs are displayed green, laws are displayed red and
the validate motif network are displayed yellow.
network col
Defines three networks (for the red, green and blue components),
change the scales as define network lum.
displ image
Choose num_image.
Choose sizes on the scales nx ny.
Click on validate displ image to copy this image in memory (available for
painting).
choose brush
click on num_brush makes this brush the current one.
win
choose the number of the image to been isplayed.
draw a frame, the inside image is stocked and can be used in
define brush type win.
line
choose the number of the image to been isplayed.
raw a line, then click on validate line: the inside image is stocked
and can be used in
define brush type line.
One can change:
dist: minimum capture distance.
adjust: number of points added on each segment.
ext: ext coefficient.
rem: removes the latest point.
contour
contour extractiuon:
choose the number of the image to been isplayed.
change coe and smooth.
Click on save to save in image number num.
smooth
image smooth:
choose the number of the image to been isplayed.
change smooth.
Click on save to save in image number num.
transp
transp image(num2)=t:
choose image1 (behind), num2 (transparent).
change transp.
Click on save to save in image number num.
medallion
image(num1,num2)center(x,y)radius(rx,ry)transp(t1,t2):
choose image1 (behind), num2 (transparent).
change transparencies t1,t2, radius rx,ry, then click in the image.
Click on save to save in image number num.
paint
just left click.
menu -> validate or press V key to save image in memroy
(needed for retrieving the same image after execution of other function).
save image
choose the number of the image to been isplayed.
ype name.eee in frame name: saves the current image in file name.eee.
hide
doesnot display the textes, make : (colon) to retrieve the display.
remove image
removes all images (YES), or not (NO).
end
returns to interpreter mode.
quit
exits the programme.
Short cuts
H: help
V: validate
I: read image
B: define brush
l: define network lum
c: define network col
d: displ image
b: choose brush
p: paint
W: win
L: line
C: contour
S: smooth
T: transp
M: medallion
s: save image
:: hide
r: remove image
q: quit
Simple use
Read images
click on read image
type ciel.bmp after name, then ENTER.
click on num_ima 1: image ima/ciel.bmp is loaded in image 1.
type visa.bmp after name, then ENTER.
click on num_ima 2: image ima/visa.bmp is loaded in image 2.
Define a brush
click on define brush:
click on num_brush 1.
click on disk.
change rx and ry.
Display an image
Click on displ image:
Click on num_ima 1.
Click on validate displ image: save the image in memory.
Paint
click on paint:
the last validated image 1 is displayed,
just click left and drag the mouse.
Press V key to save image in memroy, to write on the disk
click on save image.
Define other brush
click on define bnrush:
click on num_brush 2.
click on rectangle.
change rx and ry.
click on col 2:
choose r1,g1,b1, r2,g2,b2: interpolated colors.
click on transp 2:
choose t2 = 1
click again on paint and continue drawing with the new brush.
Change the brush
click on choose brush then on num_brush.
Neural brush
click on define network col:
3 neural networks (for red, green and blue components) are displayed,
click on define brush:
click on num_brush 3.
click on network_3.
click on image_network 1
click on transp 2, t2=1.
change rx,ry
click on paint
Press v key; shows the capture and its transformation by the network.
More advanced palette
Dynamic palette
A dynamic brush associated to an
dynamic image defines a dynamic palette for which
the trace of the brush continues to live (for example with
image spring) independently.
Pratically must be:
1) Defined a dynamic image id_im, for example:
dynamic image(id_im)=1;
mass image(id_im)=1;
spring image(id_im)=.1,-.01;
2) Definde a dynamic brush of type image id_pi associated to image id_im:
brush(id_pi)image dynamic;
texture brush(id_pi)=id_im;
programmable palette
call back
palette func(n,"Fn");
execute functions in continous on the brushes.
The trajectories are avaiable for all brushes properties.
The dynamic is avaiable.
We can define genetic images or brushes populations.
They can be connected to outputs of neural networks
whose inputs can be connected to various sensors.
Example
menu -> read image
click on name (for example ciel.bmp): this image is displayed
click on num ima 1: this image is memorized as number 1
click on an another image name (for example CRY2307.bmp: this image is displayed
click on num ima 2: this image is memorized as number 2
menu -> display image
click on num_image 1: the first image is displayed
click on num_image 2: the second image is displayed
click on validate displ image 2:this image is memorized in memory
menu -> paint: draw (right click) in the image with the default brush
menu -> validate: saves in memory
menu -> display image
click on memory: this image is displayed
menu -> window
click num image: displays this image
draw a frame in the image: this windo is saved
menu -> line
click num image: displays this image
draw a line in the image
menu -> validate: the interior of this line is saved
menu -> define brush
click on num_brush 1
click on type1 rectangle
change radius rx, ry
try changing other scales
click on num_brush 2
click on type1 win
click on num_brush 3
click on type1 win
click on num_brush 4
click on type1 line
again menu -> paint: draw (right click) in the image with the last brush
menu -> validate: saves this image
you can change the brush with menu -> choose brush,
or define new brush by menu -> define brush
menu -> contour
click on num image and change the scales: contour extraction
click on save 3: saves this image in memory
menu -> save image
click on a num_image: this image is diaplayed
enter an image name after name (for example foo.bmp): saves this image in file
foo.bmp
See also: