You are on page 1of 5

610 IEEE TRANSACTIONS ON SYSTEMS, MAN, AND CYBERNETICS, VOL. 23, NO.

2, MARCHIAPRIL 1993

For a well-behaved response, altering the deadbands, I M F . and I. INTRODUCTION


O M F in such a way that the steady-state equations are not satis- A fuzzy set F , as introduced by Zadeh [l], is a class of objects X
fied, renders the response oscillatory and, at times, unstable. These along with a grade of membership function. This membership func-
equations represent sufficient but not necessary conditions for a good tion p ~ ( z )T . E -Y,assigns to each object a grade of membership
steady -state. ranging between zero and one. Zadeh [l]assigns each object a single
value. This single value combines the evidence for z E X and the
V. CONCLUSION evidence against s E X , without indicating how much there is of
In this paper, a particular three-term fuzzy controller whose rule- each. The single number tells us nothing about its accuracy.
base was derived through fuzzy addition is analyzed. The decompo- Type 2 fuzzy sets [2] have membership functions that map from
sition of the control element is not an easy task. The use of fuzzy a set X to a type 1 fuzzy set on the interval [0,1]. One such type 2
addition facilitated this decomposition because of the symmetrical sets are @-fuzzy sets, which map from X to the set of closed
features that exist in the look-up table. However, under a center of intervals in [0,1]. Such an interval-valued fuzzy set [3, pp. 13-14]
gravity defuzzification, the conventional fuzzy logic controller was is characterized by a membership function p ~ ( s ) z, E X , which
shown in [5] to be the sum of a multilevel relay and a nonlinear PI assigns to each object a grade of membership which is a continuous
controller. Thus, the analysis carried out in this paper can be utilized interval of real numbers in the range [0,1], rather than a single
to account for the effects of the multilevel relay in the system. If value. We also assign to each object a grade of membership which
the effects of the nonlinear PI controller can be delineated, then by is a subinterval of [0,1]. This subinterval keeps track of both
using superposition the overall effects of the fuzzy controller can be the favoring evidence and the opposing evidence. However, our
obtained. So, the analysis of this paper, although particular to the definition of subinterval containment is different from the definitions
case selected, is very useful for a future analysis of a fuzzy logic of containment for other interval-valued fuzzy sets. Gratten-Guiness
controller. Moreover, if a system designer wanted a quick insight [4] defined interval containment in the natural way [a, b] I [c, d ]
into the behavior of this system, the exponential-input describing iff a 2 c and b 5 d. However, such a definition does not satisfy
function (EIDF) transient solution coupled with the SIDF steady- Theorems 1 and 2 below. Ponsard [SI defined containment as a linear
state limit cycle solution surely do provide the required information. ordering, but DeMorgan’s Law therefore did not hold.
System compensation could then be designed on an analytic basis Quinlan [6] introduced two values, t ( P ) and f(P),characterizing
using the equations given in this paper, to be checked by subsequent a proposition P. t (P ) is a greatest lower bound on the probability
computer simulations. of P derived from the evidence for P and f ( P ) is a greatest lower
bound on -P derived from the evidence against P.
REFERENCES We also use a truth membership function, t ~and, false membership
W. J. M. Kickert and E. H. Mamdani, ‘‘Analysis of a fuzzy logic function f~ to record the lower bounds on p ~ These . lower bounds
controller,” Fuzzy Sets and Systems. Amsterdam: North-Holland Pub- are used to create a subinterval of [o, 11, namely [ t F ( Z ) , 1 - f ~ ( s ) ] ,
lishing Co., 1978, pp. 29-44. to generalize the ~ . P ( sof) fuzzy sets. The lower bound and upper
G. Abdelnour, C. H. Chang, F. H. Huang, and J. Y. Cheung, “Design of bound of this subinterval are f F ( z ) and 1 - f ~ ( z )respectively.
,
a fuzzy controller using input and output mapping factors,” ZEEE Trans.
Syst., Man, Cybern., vol. 21, pp. 952-960, Sept./Oct. 1991.
These interval bounds may be derived from the bounds on ~ F ( z )
- and ~ F ( s as
) follows:

-
, “Application of describing functions in the transient response
analysis of a three-term controller,” IEEE Trans. Syst.. Man, Cybern., 1) fF(S) 5 PF(2).
pp. 603-606 this issue.
A. Gelb and W. E. Van der Velde, Multiple-Input Describing Functions 2) fF(S) 5 P,(X)

and Nonlinear System Design. New York: McGraw-Hill Electronic fF(S) I 1- P F ( Z )


Sciences Series, 1968. e 1- f d z ) 2 P F ( 2 ) .
H. Ying, W. Siler, and J. Buckley, “Fuzzy control theory: A nonlinear From 1) and 2), we obtain t F ( J-) I p~ (2) I 1 - f~ (z). It is clear
case,” Int. Federation of Automat. Contr., pp. 513-520, 1990.
that the interval is a subinterval of [0,1] since 0 is a lower bound of
all membership functions and

Vague Sets
Wen-Lung Gau and Daniel J. Buehrer
11. VAGUESETS

Abstract-A vague set is a set of objects, each of which has a grade-of


membership whose value is a continuous subinterval of [0,1]. Such a set Definition Sets’
is characterized by a truth-membership function and a falsemembership Let X be a space of points (objects), with a generic element of
function. The notion of inclusion, union, intersection, and complement are - - __ - __ - hv
-y- denoted
extended to such sets, and various properties of vague sets are established. -,T.
--
A vague set in is characterized by a truth-membership
Finally, convex vague sets are introduced.
function t v and a false-membership function fv. t v ( x ) is a lower
Manuscript received May 24, 1991; revised July 25, 1992. bound on the grade of membership of z derived from the evidence for
The authors are with the Institute of Computer Science and Information En-
gineering, National Chung Cheng University, Min Hsiung, Chia-Yi, Taiwan, T , and f ~ (s) 1 is a lower bound on the negation of z derived from the
Republic of China. evidence against z. tr/ ( z ) and f l (~z ) both associate a real number in
IEEE Log Number 9206205. +
the interval [0,1] with each point in X , where t v ( x ) f v ( z ) 5 1.

0018-9472/93$03.00 0 1993 IEEE


IEEE TRANSACTIONS ON SYSTEMS, MAN, AND CYBERNETICS, VOL. 23, NO. 2, MARCHIAPRIL 1993 611

Definition 4-Containment: A vague set A is contained in the


other vague set B , A C B , if and only if

It may not at first seem obvious why we employ this definition


of containment rather than the more obvious “subinterval” relation
- ( t 4 5 f B . 1 - f4 2 1 - f ~ ) However, . the definition makes more
xo X
sense when viewed in terms of the lower bounds for the truth/false
Fig. 1. Illustration of vague set in R1. membership functions. The statement A B has the meaning that
the lower bound on the truth of A also serves as a lower bound on the
truth of B ( f 4 <_ f B ) , and the lower bound on the falsity of B serves
That is a s a l o w e r b o u n d o n t h e f a l s i t y o f d ( f s 5 f~ e l - f q 5 1 - f ~ ) .
When the truth/falsity values are restricted to be 0 or 1, the relation
f \ : s + [O. 11
-4 C B reverts to the traditional definition that s E A + r E B ,
f r . : s -+ [O. 11. and s @ B ---t s @ A.
When we view the interval as defining lower bounds for the
This approach bounds the grade of membership of s to a subinterval truth/false membership functions, the interval has a physical inter-
[f\ (s), 1 - f\ ( s ) ] of [O, 11. pretation. Other definitions of interval containment do not correspond
In other words, the exact grade of membership h c ~( s )of s may be to this “real world” meaning. For example, a linear ordering of these
unknown, but is bounded by f i ( s )5 ( s )5 1 - fl (I),where intervals as defined in [2, p. 651, tries to compare what should be
+
f b (s) f r ( 2 ) 5 1. incomparable. If the interval is interpreted in terms of the truth/false
The precision of our knowledge about s is immediately clear, with membership functions, then the linear ordering [ a , b] 5 [c,d] iff either
our uncertainty characterized by the difference 1 - f~ (s)- tr (s). b 5 d or ( b = d and a 5 c) [2, p. 651 does not satisfy a voting model
If this is small, our knowledge about s is relatively precise; if it is of evidence. For example, according to this definition, [0.4,0.7] 5
large, we know correspondingly little. If 1- fr (s)is equal to f 1 (.r ), [0.2,0.8]. However,
our knowledge about x is exact, and the theory reverts back to that of [0.4,0.7] can be interpreted as “the vote for a resolution is 4 in
fuzzy sets. If 1- ( s )and t I (s)are both equal to 1or 0, depending favor, 3 against, and 3 abstentions.”
f ~ r

on whether s does or does not belong to I’, our knowledge about s [0.2,0.8] can be interpreted as “the vote for a resolution is 2 in
is very exact and the theory reverts back to that of ordinary sets (Le., favor, 2 against, and 6 abstentions.”
sets with two-valued characteristic functions). The number in favor for [0.4,0.7] is greater than for [0.2,0.7], but
A vague set is in R1 is illustrated in Fig. 1. the linear ordering gives a “truth” to [0.4,0.7] which is less than
When X is continuous, a vague set 1- can be written as the “truth” of [0.2,0.8]. Hence, [0.4,0.7] and [0.2,0.8] are not really
comparable, and this definition is not true in the real world. For this
reason, some restrictions must be added, making the containment
relation only a partial ordering. Therefore, Definition 4 is used. Now,
When X is discrete, a vague set I’ can be written as a comparable example using Definition 4 is as follows: [0.4,0.7] 5
[ O S , 0.81, since
n
[0.4,0.7] can be interpreted as “the vote for a resolution is 4 in
I*= C
t=1
[t\ (st), 1 - fx ( s t ) ] / s L . .r, E -Y. (4) favor, 3 against, and 3 abstentions.”
[0.5,0.8] can be interpreted as “the vote for a resolution is 5 in
For example, assume that X = [I,2 , . . ..lo]. SMALL is a vague favor, 2 against, and 3 abstentions.”
Furthermore, the number in favor for [0.5,0.8] is greater than
set of X defined by
for [0.4,0.7] and the number opposed for [0.5,0.8] is less than
small = [l,1]/1 + [0.9,1]/2 + [0.6.0.8]/3 for [0.4,0.7]. Hence [0.4,0.7] and [ O S , 0.81 are comparable and
Definition 4 is true in the real world.
+
[0.3,0.5]/4 + [0.1.0.2]/5. Definition 5: Two vague sets A and B are equal, written as
A = B , if and only if rl C_ B , and B A; that is
We start with several definitions involving vague sets that are
obvious extensions of the corresponding definitions for ordinary sets t 4 = f B (9)
and fuzzy sets. 1-fA =1-fR. (10)
Definition 2: A vague set is empty if and only it its truth-
membership and false-membership functions are identically zero Definition 6-Union: The union of two vague sets A and B with
on X . respective truth-membership and false-membership functions t A , f4,
Definition 3: The complement of a vague set A is denoted by A’ t g and fB is a vague set C , written as C = A U B , whose truth-
and is defined by membership and false-membership functions are related to those of
A and B by
tA’(s) =f4(s) (5)
1-f4t(s)=l-f4(.r). (6) tc = m a x ( t 4 , t ~ ) (11)
1 - fr = max( 1 - f 1.1 - f ~ =) 1 - nlin(fA, fs). (12)
r ) 1 - f 4 / ( s=)
In the sequel, instead of writing t A i ( s )= f ~ ( and
1 - t 4 ( s ) for all s in s,we shall write more simply f 4 , = f A , A more intuitively appealing way of defining the union is the
1-fA‘ = 1-f4. following: The union of two vague sets A and B is the smallest
612 IEEE TRANSACTIONS ON SYSTEMS, MAN, AND CYBERNETICS, VOL. 23, NO. 2, MARCHIAPRIL 1993

vague set containing both =I and B. More precisely, if D is any


vague set which contains both A and B , then it also contains the
union of .A and B . Now, we show this definition is equivalent to
Definition 2.6.
Theorem I: d U B is the smallest vague set containing both -4
and B .
Proof: C as defined by (11) and (12) contains both d-I and B,
s i n c e m a x ( t 4 . t B ) 2 tl,niax(tA.ts) 2 t s , i i i a x ( l - f 4 . 1 -
f ~ 2) 1 - f4, and max(1 - f 4 . 1 - fs)2 1 - f ~ .
Furthermore, if D is any vague set containing both A and B, then
t o 2 t l , t o 2 t ~1 - , fn 2 1 - f 4 , and 1 - fn 2 1 - f ~ . Fig. 2. Illustration of the union and intersection of two vague sets in R1.
Hence
to 2 IllaX(f4.tR) = fc function of the intersection is composed of curve segments 5 and 6.
The false-membership function of the intersection is composed of
and curve segments 7 and 8.
1 - f~ 2 iiiax(1- f4.1- f ~ =)1 - fc. Notice that although vague sets are interval-valued sets, the defini-
tions of union, intersection, and complement are completely different
which implies that C D . c Q.E.D. than the definitions for interval-valued sets. For instance, the interval-
Definition 7-Intersection: The intersection of two vague sets based union would involve curve segments 3, 4, 5 , and 6.
A and B with respective truth-membership and false-membership Theorem 3: .-I C B +-+ B’ C -4‘.
functions t A , f4,t~ and f~ is a vague set C , written as C = A n B ,
whose truth-membership and false-membership functions are related
Proof: -4 B c t 4 5 t B . 1 - f, 5 1 - f~
e fs 5 f 4 . 1 - f g 5 1 - f 4 e B’ C A’. Q.E.D.
to those of A and B by
Theorem 4: d CB u .-I n B = -4 e A UB = B
tc. = iiiin(t 1. t g j (13) Proof:
1 - f r = m i n ( l - f l . l - f ~ ) = 1 - m a x ( f 4 . f ~ ) . (14)
A cB
e t 4 5tB. 1 - f 4 5 1-fB
It is easy to show that the intersection of .-I and B is the largest u tlnR = niin(f 1 . f ~=
) t~
vague set which is contained in both .-I and B . More precisely, if
D is any vague set which is contained in both A and B , then it is l - f . \ n ~ = ~ i i i i i ( l - f ~ . l - f ~l j- f=q
also contained in the intersection of .-I and B. Now, we show this f,lug= l l l a X ( f 4 . f ~ )= t~
definition is equivalent to Definition 7. 1 - f l u s = iilax(1 - f l . 1 - fs)= 1 - fR
Theorem 2: .-I n B is the largest vague set contained in both A e A ~ B = AA. U B = B . Q.E.D.
and B.
Proof: C as defined by (13) and (14) is contained in both -4
111. S O M E PROPERTIES OF UNION, INTERSECTION
and B since
AND COMPLEMENTATION
niin(t4.tB) 5 t ? . niin(t4.tg) 5 tg With the operations of union, intersection, and complementation
iiiiii(l-fl.l-fB)5l-f4 defined on vague sets as in Section 11, it is easy to extend many of
the basic identities which hold for ordinary sets and fuzzy sets. In fact,
and the vague sets defined on a domain D with the union, intersection,
niiii(1- f4.1 - f B )5 1 and complementation operators form a DeMorgan algebra over D ,
fB.
as is shown by the properties below.
Furthermore, if D is any vague set contained in both A and Property 1--Commutativity: .-I U B = B U .-I. d n B = B n A
B, then Proof: Follows quickly from commutativity of min and max.
fD 5 f4. f D 5 fg. 1 - fo 5 1 - f 4 Property 2-Associativity: -4 U ( B U C j = (.-I U B) U C. .A n
( B n c )= (.-InB)nC.
and Proof: Follows quickly from associativity of min and max.
1-fD 5 I-fB Property 3: Idempotency:.-l U -4 = .-I. A n ’4 = A.
Proof: Follows quickly from idempotence of min and max.
Hence
Property 4-Distributivity: .-UI( B n C ) = ( A U B j n (dU C).
tn 5 n i i n ( t 4 . t B ) = tc A ~ ~ B U=C( .)- l n B j u ( . A n C )
and Proof: Follows quickly from distributivity of min and max.
Property 5: -4 n 0 = 0. -4 U .X = S. whcrc 0 =
1 - fn 5 inin(1 - f l . 1 - f ~ =)1 - fr. [O.O]. and s = [l.11 Q.E.D.
which implies that D G C. Q.E.D. Proof:
The intersection and union of two vague sets in R1 are illustrated 1) t4n0 = niin(t.l,O) = D
in Fig. 2. The truth-membership function of the union is composed 1- f4m = iniii(1 - f4.0) = 0.
of curve segments 1 and 2. The false-membership function of the 2) t . u . y = niax(t4.1) = 1
union is composed of curve segments 3 and 4. The truth-membership 1 - f s u x = inas( 1 - f . 4 . 1 ) = 1. Q.E.D.
IEEE TRANSACTIONS ON SYSTEMS, MAN, AND CYBERNETICS, VOL. 23, NO. 2, MARCHiAPRIL 1993 613

fdx), lfdx) Convex vague set Non-convex vague set t4(1.2), o f = 1 - f 4 ( 1 1 j 5 1 - f 4 ( S 2 j , then I.:, E re and
A Xa.1+ ( 1 - X j.r2 E r, by the convexity of r,.
1$4&1+Cl-Xzr2, Hence
1 -
t4(X.rl + (1 - X j s z ) 2 at = t I ( s 1 ) = niin(fA(.rl).t4(s2)j
1-fA(~.?1+(1-XjSZj >of=1-f4(Xl)

= niin(1 - f A ( S l ) , 1 - f 4 ( 2 2 ) ) .
A* Conversely, if =I is convex in the sense of the second definition
and c ) ~= t l(sl),o f = 1 - fa(sl),then may be regarded as
12
11
the set of all points for which t A ( s z ) 2 t 4 ( . r l j r 1 - f~(sz)2
Fig 3 C o n v e x and non-convex v a g u e sets in E' 1 - f 4 ( . r 1 ) . In virtue of (16) and (17), every point of the form
Asl + (1- Xj.rp, 0 5 X 5 1, is also in r, and hence rnis a
convex set. Q.E.D.
Property 6-Identity: -4U0 = .-I. -4n-Y = -4, where 0 = [O,O],
Theorem 5: If A and B are convex, so is their intersection.
and I = [l.11
Proof: Let C = A n B , then
Proof:

Property 7-Absorption: .-I U ( A n B j = -4. A n (-4 U B ) = 1 - fs(X.r1 + (1- Xjs2)).

Proof: Since -4 and B are convex:


tA(X.r1 + (1- X j s 2 j 2 niin(t4(1.1).f4(.cz))

1) t4"R = Inax(t.\.tR)
1- f 4 U ~= niax(1 - f z . 1 - f B j = 1 - niin(f4.fRj
u ti ZUB)t = niin(f 1. f~ j = t l l n B /
1 - f( l U B ) i = 1 - mas(f 4 . t B j = inin( 1 - t 1 . 1 - t R j =
f 1'"B'.
2) f i n R = niiii(t I . f R j
1 - f m B = niiii(1 - f z . 1 - f B j = 1 - n i a s ( f 4 . f B j
u ti InB)' = n l a ~ ( f ,f. R ) = t I'"B'
l - f ( I n ~ )=l 1 - n i i i i ( t 1 . t ~ )= n i a x ( l - t q . 1 - t g j =
f I'UB'. Q.E.D.
Property 9-~nvolution: ( ~ ' j ' = .-I
Proof: t = f i . 1 - f 4 f = 1 - t I r u t( = t 1.1-
f ( 4 y = 1- f l . Q.E.D.
Definition 9StrongZy Convex: A vague set A is strongly convex
if for any two distinct points s1 and s 2 , and any X in the open
IV. CONVEXITY
interval (0,l)
We assume for concreteness that .Y is a real Euclidean space E".
Definition 8: Convexity: A vague set -4 is convex if and only if + (1 - X).rzj > r n i n ( t ~ ( r l j . t s ( z 2 ) ) (18)
t / L ( h 1

the sets rc, defined by 1 - fA(X.rl + (1 - j > min( 1 -


Xj1.2 f4(1.1), 1 - f 4 ( . ~ 2 ) ) . (19)
r,,= {.r l t 4 ( . r j 2 n r . i- f l ( . r j 2 n f } (15)
Theorem 6: If A-Iand B are strongly convex, so is their intersec-
are convex for all a,,n in the interval (O,1].
An alternative and more direct definition of convexity is the
following: =I is convex if and only if Proof: The proof is similar to that of Theorem 5.

for all and , r ~in X and all X in [0, I]. This is illustrated in Fig. 3 from ordinary sets, fuzzy sets, and interval-valued sets. The notions of
for 7) = 1. inclusion, union, intersection, complement, relation, and composition
To show the equivalence between the above definitions, note that if have been extended to vague sets. Finally, convex vague sets and
-4 is convex in the sense of the first definition, and nf = f 4 ( s1) 5 strongly convex sets have been introduced.
614 IEEE TRANSACTIONS ON SYSTEMS, MAN, AND CYBERNETICS, VOL. 23, NO. 2, MARCHIAPRIL 1993

The major advantage of vague sets over fuzzy sets is that vague I. INTRODUCTION
sets separate the positive and negative evidence for membership of Few problems in computational vision have drawn more attention
an element in the set. We not only have an estimate of how likely it than the representation of surfaces. A desirable property of any
is that an element is in the set, but we also have an lower and upper surface representation is that it be invariant with respect to rigid trans-
bound on this likelihood. This lower/upper bound can be used to formations of the surface in 3-space, otherwise one cannot guarantee
perform constraint propagation. For instance, constraint propagation the integrity of the representation under changes in viewpoint. This
can be used to detect inconsistencies in assignments of intervals to paper considers the computation of invariant surface representations
Boolean expressions involving the sets. given 3-D depth information from single and multiple views. Our
We are currently using constraint propagation to generalize the approach makes use of an elastically deformable model that offers the
methods of first-order automatic theorem proving so as to find a geometric flexibility to satisfy a plethora of multimodal and multiview
fixed-point minimal-sized possibility interval for each predicate in constraints. We specialize this model to assume an initial cylindrical
the input clauses, or to find inconsistencies, if they exist [7]. shape as in Terzopoulos et al. [21]. We dub the process of surface
fitting with elastically deformable surface as dynamic surface fitting.
REFERENCES In dynamic surface fitting, all the data are transformed into externally
L. A. Zadeh, “Fuzzy sets,” Inform. Control, vol. 8, pp. 338-353, 1965. applied forces. Constraints from single as well as multiple views are
D. Dubois and H. Prade, Fuzzy Sets andsystems: Theory andilpplications. applied as external forces to mold the elastic tube model into a shape
New York: Academic, 1980. consistent with the data.
G. J. Klir and T. A. Folger, Fuzzy Sets, Uncertainty, and Information. Surface reconstruction typically precedes the symbolic represen-
Englewood Cliffs, NJ: Prentice Hall, 1988.
I. Grattan-Guiness,“Fuzzy membership mapped onto intervals and many-
tation of surfaces. The reconstruction process integrates raw shape
valued quantities,” Zeitschr. Math. Logik. and Grundlagen d. Math., data from multiple visual sources. Using this information it generates
Bd. 22, S, pp. 149-160, 1975. a dense, coherent surface. Several variational reconstruction methods
C. Ponsard, “Hierarchie des places centrales et graphes +-flous,” Environ. have been proposed. An efficient noninvariant method was developed
Plann. A 9, pp. 1233-1252, 1977.
by Terzopoulos [19]. Blake [2] pointed out that the noninvariance
J. R. Quinlan, “Inferno: A cautious approach to uncertain inference,”
Computer J., vol. 26, no. 3, pp. 255-268, 1983. was inherent to the functional used in the variational principle
D. J. Buehrer, “Vague logic: A first-order extension of interval probability formulation, which quantifies the small deflection bending energy of
theory,” in Proc. IEEE Workshop Imprecise and Approximate Computa- a thin flexible plate. He proposed an invariant functional involving
tion, Dec. 1, 1992, Phoenix, AZ, pp. 83-87. the principal curvatures (minimum and maximum normal curvatures)
of the surface. This is known as the large deflection bending energy
of a thin plate.
A problem with Blake’s approach is that the large deflection energy
is a nonquadratic functional that, moreover, is nonconvex. The result
Intrinsic Parameters for Surface Representation is a difficult, nonlinear system of equations with multiple suboptimal
Using Deformable Models solutions. Blake and Zisserman [3] use a deterministic method for
finding the global optimum, which they claim compares favorably
Baba C. Vemuri and Ravikanth Malladi against Monte Carlo techniques such as simulated annealing [ 111.
Like ofher methods for solving large nonlinear systems, however,
the computational complexity of their algorithm is equivalent to the
Abstruct- Three-dimensional viewpoint invariance is an important solution of many linear systems of comparable size. By contrast,
requirement on the representation of surfaces for recognition tasks. noninvariant reconstruction as in [ 191 uses a quadratic functional,
Parameterized surfaces possess this desirable property. In general, the
parameters in a parametric surface representation can be arbitrarily and finding the unique minimum requires the solution of a single
defined. A canonical, intrinsic parameterization provides a consistent, linear system.
invariant form for describing surfaces. Our goal here is to define and In this paper, we develop a viewpoint invariant surface repre-
construct such a parameterization. A new technique for achieving this sentation scheme whose computational complexity is closer to that
goal is presented by using an elastically deformable model. The salient
features of our method are that it provides a unified and general of noninvariant reconstruction. To achieve our goal of invariance,
framework for reparameterization of a surface and easily allows for however, we must dissociate the surface description from the viewer
incorporation of multiview data sets. The canonical parameterization of centered coordinate frame. This can be done easily by employing an
the surface is defined in terms of the surface lines of curvature. Depth intrinsically parameterized surface energy functional in the variational
constraints are first imposed as an external force field on the deformable
model that molds itself to be consistent with the data. Principal vectors reconstruction. We adopt an intrinsic parameterization for the surface
computed from this conformed model surface are then imposed as a that depends purely upon the surface shape and is thereby invariant
force field on the parameter curves of the model. The parameter curves with respect to rigid motion. We express the resultant surface in a
deform to become tangential to the principal vectors thereby yielding an parametric form, where the parameters are curves in predetermined
invariant surface parameterized by the lines of curvature. Extension of
the canonical parametric grid to multiple views is then demonstrated by
directions along the surface [24]-[26], [28]. To obtain a unique
incorporating depth and curvature constraints from multiple views. parameterization, we choose these directions to be the principal
directions, which correspond to the minimum and maximum normal
Manuscript received April 21, 1991; revised July 18, 1992. This work was curvatures of the surface. These directions by Euler’s theorem [9]
supported in part by National Science Foundation under the contract No. EET- are mutually orthogonal. Our final representation is therefore an
8810751.
The author is with the Department of Computer & Information Sciences, orthogonally gridded surface where the grid coordinates are along the
University of Florida, Gainesville, FL 32611. integral curves of the principal vector fields. We call this a canonical
IEEE Log Number 9206200. representation. Orthogonal grids are useful in applications involving

0018-9472/93$03.00 0 1993 IEEE

1 - -

You might also like