validate anc
Recalculates the anchor matrices.
validate field(id)
Returns the field id validation.
validate field(id)=v
Changes this validation.
Notes:
1) v=1: active (default).
v=0: passive.
2) all fields by default.
validate col
validate(1)col vertex vol(id)
Saves col vertex vol(id) in block
col dup vertex of volume id.
validate(-1)col vertex vol(id)
Restores col vertex vol(id).
validate fac
validate fac(f) vol(id)
validate fac
validate fac(f) vol(id)
Returns the facet f of volume id validation.
validate fac(f) vol(id)=v
Validates the facet f of the volume id.
v=-1: inhibits facet f (not displayed).
v=1: validates facet f (default).
All facets by default.
validate fac(f)ext vol(id)
Returns the extension facet f of volume id validation.
validate fac(f)ext vol(id)=v
Validates the extension facet f of the volume id.
validate genetic
validate validate genetic(id)
Returns the validation flag of genetic(id) and NIL
if tne genetic does not exist.
validate validate genetic(id)=v
Change this flag.
validate genetic(id)
Evaluates genetic id (all by default) that is calculated in
scale particle genetic(id) the returns of the
evaluation function for each individual of the population id.
Options:
particle(num): evluates only particles num.
var: returns the average evaluatiuon of the genetic (or of the particles num
if option particle(num) is present).
validate image
validate image(id)
Returns the validation flag of the image id.
validate image(id)=v
Changes this flag.
Note:
v = 0: invalid, image not displayed (default).
v = 1: valid, image displayed.
v = -1: image not displayed but possibly animated.
validate(1)image(id)
Saves the image id in the block validate.
validate(-1)image(id)
Restores the image id if it has been saved.
Note: not be confused with validate image(id)=v.
validate(v)image image(id)
v > 0: saves bloc image in bloc -image
v < 0: restaures bloc -image in bloc image
validate interaction
validate interaction write name("nnn")
Returns 0 if animation (initiated by interaction write name'"nnn")
is completed.
validate interaction ascii("c")func
Returns 1 if c key have been pressed 1, 3, 5, ...times,
and returns 0 if c key have been pressed 0, 2, 4, ... times.
validate interaction ascii("c")func=val: changes this value.
validate inv law network
validate inv law(L) network(r1,r2,...)
Determines the network ri which a law was closest to L and returns the motif correcpondant.
Useful in
environement multi networks.
validate law CG vol(id)
Returns the power 1/3 law evaluation of the law CG vol(id).
validate limit
validate(x)limit(x1,x2,y1,y2)
Returns y = y1 + (x - x1) * (y2 - y1) / (x2 - x1)
validate(x)limit(x1,x2,y1,y2)inv
Returns x = x1 + (y - y1) * (x2 - x1) / (y2 - y1)
Note:
Used to code motif and decode law of a neural network.
Examples:
motif(0)network(1)=validate(.25*PI)limit(-P1,PI,0,1);
codes an angle.
an=validate(0.625)limit(-P1,PI,0,1)inv;
decodes an angle.
motif(0)network(1)=validate(-1,0,1)limit(-1,1,0,1);
codes an axis.
ax=validate(0,.5,1)limit(-1,1,0,1)inv;
decodes an axis.
validate local
validate local obj(id)
Returns the validate flag of the 1st local function
of the object id of type obj.
validate local obj(id)=b
Changes this flag.
Notes:
All objects of type obj by default.
This command allows you to ignore (validate = 0) or make it active (validate = 1) local functions of the object id.
validate menu
validate menu
Returns the validation flags of menus..
validate menu=v
v=1: visible menus (default) equivalent to yes menu.
v=0: invisible menus equivalent to no menu.
validate menu("MEN")
Returns the validation flag of the menu "MEN"
validate menu("MEN")=v
v=1: menu "MEN" visible (default).
v=0: menu "MEN" invisible
.
validate menu("f0")menu("fi")
Returns the validation flag of the function fi of the menu f0.
(>0 if active, <0 if passive, 0 si neutral).
validate menu("f0")menu("fi")=1
Makes active function fi of the menu f0.
validate menu("f0")menu("fi")=0
Makes passive function fi of menu f0.
Note:
validate menu("f0")menu("*")=v
Change the flag of all functions of menu f0.
validate menu interaction
Regenerates the interaction menus.
validate message
validate message(id)
Returns the validation flag of the message identifier id.
validate message(id)=v
Changes this flag.
v=1: active message (default).
v=0: inhibited message.
validate motif
validate motif(in) network(id)
Returns the output of the neural network id for the input in.
validate motif(in) network(id)speed
Returns the output speed of the id for network in.
validate motif(m)network(1)neuronne
Returns the number of the winner neuron and its activation
when the network is of type Kohonen.
validate motif(m) network(r1,r2,...)
Determines, the network ri, for which a motif is nearest of m and do:
validate motif(m) network(ri), returning the correspondant law.
validate motif(m) average network(r1,r2,...)
Determines, for each network ri, the minimum distance di from m to all learning motifs,
and returns the weighted (by di) sum of outputs of these networks for input m.
Useful in environement of multi networks.
See also:
validate network
validate inv law network
validate motif(m)network(id)roll
Returns the output of the network id for input m, the motifs block works like a circular stack
in which is inserted m (useful fornetworks of type near).
validate normal
validate normal vol(id)
If the vertices of the volume id have no normal, they are created null (0,0,0).
validate obj
validate obj(id)
Returns the validation flag of the object id of type obj:
1 validated.
0 invalidated (inactive and not viewable, but its followers are displayed).
-1 if object is invalidated as well as its followers.
-2 if object is invalidated but not its motifs.
-3 if object is invalidated as well as its followers, but not its motifs.
-4 inhibites the dynamic for this object.
-5 taken into account but not displayed;.
NIL if object does not exist.
Notes:
1) validate obj=v; validated (v=1), or invalidated (v=0) all objects of type obj.
2) For a network validate network has another meaning, use
validate validate network(id).
NP motif network(id)
returns NIL if network id does not exist, 0 if it exists but does not has motif, otherwise returns
the number of the motifs.
3) For a genetic validate genetic has another meaning, use
validate validate genetic(id).
NP particle genetic(id) returns NIL if genetic id does not exist,
0 if it exists but has no particle, otherwise returne the number of the particles.
4) For a volumes structure anchored id0,id1,id2,... of root id0,
if you want to view only the volume idi, write:
validate vol(id0,id1,id2,...)=0;validate vol(id0,idi)=1;
validate obj(id)=v
Changes this flag.
Note:
v = 1: object is active.
v = 0: object is inactive.
validate obj: all objects obj.
validate palette
Validates the palette mode: useful before display.
validate particle
validate particle(p) genetic(id)
Evaluates the individual p of the genetic id, ie computes in
scale particle(p) genetic(g) the return of the
evaluation function func genetic(g) for individual
p of the population id.
validate pers
validate pers(x,y,z)
In a local function, this cammand validates the perspective at point (x,y,z).
Notes:
See procedural perspective.
Must do: yes local and yes pers
validate pers vol(id)
The algorithm implementation classical perspective (conical) using
homogeneous coordinates (4D).
In principle it is up to deform the database as if it were cylindrical perspective (from an eye to infinity).
The advantage of this method is to reduce the conical projection to a cylindrical projection (easier especially with regard to the clipping).
The validate pers vol actually generates the modified database. For another point of view we get strange effects ...
pers validate vertex (s) vol (id)
Only the vertices numbered s of thevolume id will be treated.
validate pixel
validate(1) pixel image(id)
Saves the image id pixels.
validate(-1) pixel image(id)
Restores the image id pixels.
validate poi
validate(1) poi vol(id)
Saves the volume id points in thevalidate poi block.
validate(-1) poi vol(id)
Restores the volume id points.
Options:
poi(n): restores only points number n.
frac ext or generate: processes with the extension points
or with the fractal points.
Note: if the validate poi block does nt exist, validate(-1) poi vol(id)
builds it by duplication of poi block.
validate properties
validate propr1 propr2 vol(id)
Transmits as inputs of the network network propr1 propr2 vol(id)
parameters of the property propr1, the output provides the parameters boundaries of the property propr2
which then forced inside this interval.
Note: First must be
built networks by:
r=network propr1 propr2 vol(id)
defined learning pairs:
motif propr1 propr2 vol(id)=M1,M2,....
law propr1 propr2 vol(id)=L1,L2,....
and trained the networks by:
validate(n)network(r).
validate network
validate(n,i0,nc) network(id)
Adapts the network id weights to the set of pairs (motif, law) in n passes maximum.
If i0 is absent or i0=0: nb evaluations.
Otherwise if i0 > ;0: parallel processing image i0
If nc is present: depth of the recurrence (network near).
Returns cpt,err,stat
cpt = number of steps actually realized
err = maximum error committed
stat = 1 if end of a parallel processing
validate(n,i0,nc)error(max_err)coe(c1,c2)time(t)network(id)
Allows specify some parameters:
1) Learning is stopped as soon as the error is less than max_err.
2) c1 and c2 are the
learning constants.
(1.0, 0.01 default).
time(t) for network of type
near.
validate network number
Returns the network number ri which a pattern is closest to m, see:
validate motif(m)network(r1,r2,..)
Example of use (in a program loop):
nb=validate(n,1)network(1)error(.02)read("r");
nb=validate(n,2)network(1)error(.02);
....................................
nb=validate(n,n)network(1)error(.02);
if(nb[2]): end
validate(n) coe(c1,c2) error(eps) network(num)
The "eta" value of
errors correction varies from c1 (step 1)
to c2 (step n). By default c1=c2=0.5
validate(n) law(0,c1,0,.1,c2,0,...,1,cn,0) error(eps) network(num)
The "eta" value of the errors correction varies according a law
Example:
validate(5000) law(0,1,0, .5,.2,0, 1,.1,0) network(1)
Learning stops as soon as the error is less than eps
by default eps=0.1
validate(n) network(num) edit
Couples (output-desired, output-calculated) are published
Errors are highlighted by: ********
The errors number and the maximum error are published.
The steps number and the maximum error are returned
see
network
validate(n) network(id) read("F")
F is a file of the form:
x1,x2,...,xn
y1,y2,...,ym
x1,x2,...,xn
y1,y2,...,ym
....
Inputs motifs are: (x1,x2,...,xn)
The network id is created ans trained in n steps.
Options:
write("R"): the adapted network is saved in file R.res
mass(m1,m2): weights are randomly generated between m1 and m2.
(-1,1 defult)
rand: no reproductible random (default)
alea: reproductible random
alea(ind): reproductible of seed num
error(eps): the tolerated error is eps (0.1 par défaut)
validate(n) network(num) read("F") format(m1,m2,l1,l2)
Default motifs and laws must be normalized between 0 and 1. format (m1, m2, l1, l2) can return motifs
between m1 and m2, and laws between l1 and l2.
See also:
validate motif network
validate inv law network
validate network vol
validate(n,i0,nc) network vol(id)
Built an internal network with 7 neurons to volume id:
(1,2,3) the input layer receiving the axes (x, y, z).
(4.5) of the hidden layer.
(6.7) of the output layer outputting boundaries (a1, a2).
The network is trained on n passes.
validate network vol(id)
At each image of the animation (typically a function func func(4,"F4")
passes the axis (x, y, z) of the volume id as entry of the network which provides an output (a1, a2),
the rotation angle a of the volume is forced id between a1 and a2.
validate validate network(id)
Returns the validation flag of network id and NIL if tne network does not exist.
validate validate network(id)=v
Change this flag.
validate mouse
validate mouse
Returns the 3D polygonal line entry by mouse.
validate mouse speed
Returns the 3D polygonal line entry by mouse speed
(x1,y1,0,x2,y2,t1,...,xn,yn,tn) with ti = input time.
validate see
validate see
Switches in see mode: see bar appears at the top right corner of the screen,
if you click on the dash is activated see.
validate(0)see
Inhibits see mode.
validate speed
validate speed vol(id)
If the volume has no mass, or if no dynamic is set,
the speed of volume id is computed.
Options:
validate(v): multiplier of speed (1 default).
validate speed rota vol(id)
If the volume has no mass, or if no dynamic is set,
the speed rota of volume id is computed.
Options:
validate(v): multiplier of speed (1 default).
validate sphere
validate sphere
Returns the display flag of the object sphere.
validate sphere=v
Gives value v to the display flag of the object sphere.
v=1: displayable.
v=0: no displayable.
validate view
validate view(id)
Returns the view id flag (all views by default).
validate view(id)=v
Gives to view id flag the v value:
v=1: active.
b=0: passive.
All views by default.
Notes:
For several activated views, must do yes view
validate win
validate win(id)
Returns the window id validation.
validate win(id)=v
Validates (v=1) or inhibits (v=0) the window id.
See also: