The term pre<strong>shape</strong> space is sometimes used for the leftmost space, when bothspaces are studied <strong>in</strong> the same paper.4.5 Geometric <strong>curves</strong>Unfortunately the quotientB i = Imm(S 1 , lR n )/Diff(S 1 )<strong>of</strong> immersed <strong>curves</strong> up to reparameterization is not a Fréchet manifold.We (re)def<strong>in</strong>e the space <strong>of</strong> geometric <strong>curves</strong>.Def<strong>in</strong>ition 4.9B i,f (S 1 , lR n ) = Imm f (S 1 , lR n )/Diff(S 1 )is the quotient <strong>of</strong> Imm f (S 1 , lR n ) (the free immersions) by the diffeomorphismsDiff(S 1 ) (that act as reparameterizations).The good news is thatProposition 4.10 (§2.4.3 <strong>in</strong> Michor <strong>and</strong> Mumford [37]) If Imm f has thetopology <strong>of</strong> the Fréchet space <strong>of</strong> C ∞ functions, then B i,f is a Fréchet manifoldmodeled on C ∞ .The bad news is that• when we add a simple Riemannian metric to B i,f , the result<strong>in</strong>g metricspace is not metrically complete; <strong>in</strong>deed, there cannot be any norm on C ∞that generates the same topology <strong>of</strong> the Fréchet space C ∞ (as we discussed<strong>in</strong> 3.26);• by model<strong>in</strong>g B i,f as a Fréchet manifold, some calculus is lost, as we saw <strong>in</strong>Section 3.2.5.Remark 4.11 It seems that this is the only way to properly def<strong>in</strong>e the manifold.If otherwise we choose M = C k (S 1 → lR n ) to be the manifold <strong>of</strong> <strong>curves</strong>, then ifc ∈ M, c ′ ∉ T c M. Hence we (must?) model M on C ∞ functions.4.5.1 Research pathFollow<strong>in</strong>g Michor <strong>and</strong> Mumford [37] we so obta<strong>in</strong>ed a possible program <strong>of</strong> mathresearch:• def<strong>in</strong>eB = B i,f (S 1 , lR n ) = Imm f (S 1 , lR n )/Diff(S 1 )<strong>and</strong> consider B as a Fréchet manifold modeled on C ∞ ,• def<strong>in</strong>e a Riemann/F<strong>in</strong>sler geometry on it, study its properties,• metrically complete the space.In the last step, we would hope to obta<strong>in</strong> a differentiable manifold; unfortunately,this is sometimes not true, as we will see <strong>in</strong> the overview <strong>of</strong> the literature.38
4.6 Goals (revisited)We formulate an abstract set <strong>of</strong> goal properties on a metric 〈h 1 , h 2 〉 G|c on spaces<strong>of</strong> <strong>curves</strong>.1. [rescal<strong>in</strong>g <strong>in</strong>variance] For any λ > 0, if we rescale c to λc, then〈h 1 , h 2 〉 G|λc = λ a 〈h 1 , h 2 〉 G|c(where a ∈ lR is an universal constant);2. [Euclidean <strong>in</strong>variance] Suppose that A is an Euclidean transformation,<strong>and</strong> R is its rotational part; if we apply A to c <strong>and</strong> R to h 0 , h 1 , then〈Rh 1 , Rh 2 〉 G|Ac = 〈h 1 , h 2 〉 G|c ;3. [parameterization <strong>in</strong>variance]the metric does not depend on the parameterization <strong>of</strong> the curve, that is‖˜h‖˜c = ‖h‖ c when ˜c(t) = c(ϕ(t)) <strong>and</strong> ˜h(t) = h(ϕ(t)).If a metric satisfies the above three properties, we say that it is a geometricmetric.4.6.1 Well posednessWe also add a more basic set <strong>of</strong> requirements.0. [well-posedness <strong>and</strong> existence <strong>of</strong> m<strong>in</strong>imal geodesics]• The metric <strong>in</strong>duces a good distance d: that is, the distance between different<strong>curves</strong> is positive, <strong>and</strong> d generates the same topology that the atlas <strong>of</strong> themanifold M <strong>in</strong>duces;• (M, d) is complete;• for any two <strong>curves</strong> <strong>in</strong> M, there exists a m<strong>in</strong>imal geodesic connect<strong>in</strong>g them.4.6.2 Are the goal properties consistent?So we state the abstract problem:Problem 4.12 Consider the space <strong>of</strong> <strong>curves</strong> M, <strong>and</strong> the family <strong>of</strong> all Riemannian(or, F<strong>in</strong>sler) Geometries F on M.Does there exist a metric F satisfy<strong>in</strong>g the above properties 0,1,2,3?Consider metrics F that may be written <strong>in</strong> <strong>in</strong>tegral form∫F (c, h) = f ( c(s), ∂ s c(s), . . . , ∂ j sc(s), h(s), . . . , ∂ i j s ih(s)) dscwhat is the relationship between the degrees i, j <strong>and</strong> the properties <strong>in</strong> this section?All this boils down to a fundamental question: can we design metrics to satisfyour needs?39
- Page 1 and 2: Metrics of curves in shape optimiza
- Page 3 and 4: shape analysis where we study a fam
- Page 5 and 6: • F (c) = F (c ◦ φ) for all cu
- Page 7 and 8: κ > 0HNNHNHκ < 0Figure 1: Example
- Page 9 and 10: In the case of planar curves c 1 ,
- Page 11 and 12: (a) (b) (c) (d) (e)Figure 3: Segmen
- Page 13 and 14: where φ may be chosen to beφ(x) =
- Page 15 and 16: 2.4.1 Example: geometric heat flowW
- Page 17 and 18: 2.4.5 Centroid energyWe will now pr
- Page 20 and 21: shapes. Unfortunately, H 0 does not
- Page 22 and 23: • If the second request is waived
- Page 24 and 25: • The Fréchet space of smooth fu
- Page 26 and 27: Since φ k are homeomorphisms, then
- Page 28 and 29: 3.6.1 Riemann metric, lengthDefinit
- Page 30 and 31: Theorem 3.30 Suppose that M is a sm
- Page 32 and 33: Example 3.38 Let M = C ∞ ([−1,
- Page 34 and 35: • sometimes S 1 will be identifie
- Page 36 and 37: The proof is by direct computation.
- Page 40 and 41: 5 Representation/embedding/quotient
- Page 42 and 43: 6.1.1 Length induced by a distanceI
- Page 44 and 45: 0101200000000000011111111111110000
- Page 46 and 47: 6.2.4 Applications in computer visi
- Page 48 and 49: of a small ball from A. The motion
- Page 50 and 51: CutΩ(The arrows represent the dist
- Page 52 and 53: 7.3 L 1 metric and Plateau problemI
- Page 54 and 55: Definition 8.3 (Flat curves) Let Z
- Page 56 and 57: Proof. Fix α 0 ∈ S \ Z. Let T =
- Page 58 and 59: 8.2.2 RepresentationThe Stiefel man
- Page 60 and 61: 9.3 Conformal metricsYezzi and Menn
- Page 62 and 63: 10.1.1 Related worksA family of met
- Page 64 and 65: Definition 10.7 (Convolution) A arc
- Page 66 and 67: and̂∇˜HjE(0) = ̂∇ H 0E(0),̂
- Page 68 and 69: and we simply integrate twice! More
- Page 70 and 71: Corollary 10.18 In particular, the
- Page 72 and 73: 10.6 Existence of gradient flowsWe
- Page 74 and 75: • The length functional (from C 1
- Page 76 and 77: We can eventually estimate the diff
- Page 78 and 79: 7.552.5-10 -7.5 -5 -2.5 2.5 5 7.5 1
- Page 80 and 81: By using (i) and (ii) from lemma 10
- Page 82 and 83: 10.6.3 Existence of flow for geodes
- Page 84 and 85: 10.7.1 Robustness w.r.to local mini
- Page 86 and 87: We already presented all the calcul
- Page 88 and 89:
10.9 New regularization methodsTypi
- Page 90 and 91:
can be solved for k (this is not so
- Page 92 and 93:
2. Now• at t = 1/2 it achieves th
- Page 94 and 95:
Definition 11.10 • The orbit is O
- Page 96 and 97:
y eqn. (11.3), where the terms RHS
- Page 98 and 99:
So Prop. 11.19 guarantees that the
- Page 100 and 101:
Theorem 11.23 Suppose that the Riem
- Page 102 and 103:
2.2.3 Examples of geometric energy
- Page 104 and 105:
9 Riemannian metrics of immersed cu
- Page 106 and 107:
contour continuation, 11convolution
- Page 108 and 109:
normal vector, 6objective function,
- Page 110 and 111:
References[1] Luigi Ambrosio, Giuse
- Page 112 and 113:
[30] Serge Lang. Fundamentals of di
- Page 114 and 115:
[57] Ganesh Sundaramoorthi, Anthony