Powell--Sabin splines on the sphere with applications in...

61
Powell–Sabin splines Spherical Powell–Sabin splines Multiresolution analysis Powell–Sabin splines on the sphere with applications in CAGD Jan Maes Department of Computer Science Katholieke Universiteit Leuven Paris, November 17, 2006

Transcript of Powell--Sabin splines on the sphere with applications in...

  • Powell–Sabin splines Spherical Powell–Sabin splines Multiresolution analysis

    Powell–Sabin splines on the sphere withapplications in CAGD

    Jan Maes

    Department of Computer ScienceKatholieke Universiteit Leuven

    Paris, November 17, 2006

  • Powell–Sabin splines Spherical Powell–Sabin splines Multiresolution analysis

    Outline

    Section I Powell–Sabin splines

    Section II Spherical Powell–Sabin splines

    Section III Multiresolution Analysis

  • Powell–Sabin splines Spherical Powell–Sabin splines Multiresolution analysis

    Powell–Sabin splines

  • Powell–Sabin splines Spherical Powell–Sabin splines Multiresolution analysis

    Bernstein–Bézier representation

    =⇒

    Pierre Étienne Bézier (1910-1999)

  • Powell–Sabin splines Spherical Powell–Sabin splines Multiresolution analysis

    Stitching together Bézier triangles

    =⇒

    No C1 continuity at the red curve

  • Powell–Sabin splines Spherical Powell–Sabin splines Multiresolution analysis

    C1 continuity with Powell–Sabin splines

    Conformal triangulation ∆

    PS 6-split ∆PS

    S12(∆PS) = space of PS splines

    M.J.D. Powell M.A. Sabin

  • Powell–Sabin splines Spherical Powell–Sabin splines Multiresolution analysis

    C1 continuity with Powell–Sabin splines

    Conformal triangulation ∆

    PS 6-split ∆PS

    S12(∆PS) = space of PS splines

    M.J.D. Powell M.A. Sabin

  • Powell–Sabin splines Spherical Powell–Sabin splines Multiresolution analysis

    C1 continuity with Powell–Sabin splines

    Conformal triangulation ∆

    PS 6-split ∆PS

    S12(∆PS) = space of PS splines

    M.J.D. Powell M.A. Sabin

  • Powell–Sabin splines Spherical Powell–Sabin splines Multiresolution analysis

    The dimension of S12(∆PS)?

    There is exactly one solution s ∈ S12(∆PS) to theHermite interpolation problem

    s(Vi) = αi ,

    Dxs(Vi) = βi , ∀Vi ∈ ∆, i = 1, . . . ,N.Dys(Vi) = γi ,

    The dimension of S12(∆PS) is 3N. Therefore we need 3N basis

    functions.

  • Powell–Sabin splines Spherical Powell–Sabin splines Multiresolution analysis

    The dimension of S12(∆PS)?

    There is exactly one solution s ∈ S12(∆PS) to theHermite interpolation problem

    s(Vi) = αi ,

    Dxs(Vi) = βi , ∀Vi ∈ ∆, i = 1, . . . ,N.Dys(Vi) = γi ,

    The dimension of S12(∆PS) is 3N. Therefore we need 3N basis

    functions.

  • Powell–Sabin splines Spherical Powell–Sabin splines Multiresolution analysis

    Powell–Sabin B-splines with control triangles

    s(x , y) =N∑

    i=1

    3∑j=1

    cijBij(x , y)

    Bij is the unique solution to

    [Bij(Vk ),DxBij(Vk ),DyBij(Vk )] = [0,0,0] for all k 6= i[Bij(Vi),DxBij(Vi),DyBij(Vi)] = [αij , βij , γij ] for j = 1,2,3

    Partition of unity:∑Ni=1

    ∑3j=1 Bij(x , y) = 1,

    Bij(x , y) ≥ 0

    (Paul Dierckx, 1997)

  • Powell–Sabin splines Spherical Powell–Sabin splines Multiresolution analysis

    Powell–Sabin B-splines with control triangles

    s(x , y) =N∑

    i=1

    3∑j=1

    cijBij(x , y)

    Bij is the unique solution to

    [Bij(Vk ),DxBij(Vk ),DyBij(Vk )] = [0,0,0] for all k 6= i[Bij(Vi),DxBij(Vi),DyBij(Vi)] = [αij , βij , γij ] for j = 1,2,3

    Partition of unity:∑Ni=1

    ∑3j=1 Bij(x , y) = 1,

    Bij(x , y) ≥ 0

    (Paul Dierckx, 1997)

  • Powell–Sabin splines Spherical Powell–Sabin splines Multiresolution analysis

    Powell–Sabin B-splines with control triangles

    s(x , y) =N∑

    i=1

    3∑j=1

    cijBij(x , y)

    Bij is the unique solution to

    [Bij(Vk ),DxBij(Vk ),DyBij(Vk )] = [0,0,0] for all k 6= i[Bij(Vi),DxBij(Vi),DyBij(Vi)] = [αij , βij , γij ] for j = 1,2,3

    Partition of unity:∑Ni=1

    ∑3j=1 Bij(x , y) = 1,

    Bij(x , y) ≥ 0

    (Paul Dierckx, 1997)

  • Powell–Sabin splines Spherical Powell–Sabin splines Multiresolution analysis

    Powell–Sabin B-splines with control triangles

    Three locally supported basis functions per vertex

  • Powell–Sabin splines Spherical Powell–Sabin splines Multiresolution analysis

    Powell–Sabin B-splines with control triangles

    The control triangle is tangent to the PS spline surface.

  • Powell–Sabin splines Spherical Powell–Sabin splines Multiresolution analysis

    Powell–Sabin B-splines with control triangles

    It ‘controls’ the local shape of the spline surface.

  • Powell–Sabin splines Spherical Powell–Sabin splines Multiresolution analysis

    Spherical Powell–Sabin splines

  • Powell–Sabin splines Spherical Powell–Sabin splines Multiresolution analysis

    Spherical spline spaces

    P. Alfeld, M. Neamtu, and L. L. Schumaker (1996)

    Homogeneous of degree d : f (αv) = αd f (v)Hd := space of trivariate polynomials of degree d that arehomogeneous of degree dRestriction of Hd to a plane in R3 \ {0}⇒ we recover the space of bivariate polynomials∆ := conforming spherical triangulation of the unit sphere S

    Srd(∆) := {s ∈ Cr (S) | s|τ ∈ Hd(τ), τ ∈ ∆}

  • Powell–Sabin splines Spherical Powell–Sabin splines Multiresolution analysis

    Spherical Powell–Sabin splines

    s(vi) = fi , Dgi s(vi) = fgi , Dhi s(vi) = fhi , ∀vi ∈ ∆

    has a unique solution in S12(∆PS)

  • Powell–Sabin splines Spherical Powell–Sabin splines Multiresolution analysis

    1− 1 connection with bivariate PS splines

    ⇒ |v |2Bij(v|v |

    )⇒

    ←−

    Spherical PS B-spline Bij(v)

    piecewise trivari-ate polynomial ofdegree 2 that ishomogeneous ofdegree 2

    Restriction to theplane tangent toS at vi ∈ ∆

  • Powell–Sabin splines Spherical Powell–Sabin splines Multiresolution analysis

    1− 1 connection with bivariate PS splines

    Let Ti be the plane tangent to S at vertex viRadial projection:

    Riv := v :=v|v |∈ S, v ∈ Ti

    Define ∆i as the star of vi in ∆, and let ∆PSi ⊂ ∆PS be its

    PS 6-split.

    Theorem

    Let s ∈ S12(∆PSi ). Let s be the restriction of |v |2s(v/|v |) to Ti .

    Then s is in S12(R−1i ∆

    PSi ) and

    s(vi) = s(vi), Dgi s(vi) = Dgi s(vi), Dhi s(vi) = Dhi s(vi).

  • Powell–Sabin splines Spherical Powell–Sabin splines Multiresolution analysis

    1− 1 connection with bivariate PS splines

    Let Ti be the plane tangent to S at vertex viRadial projection:

    Riv := v :=v|v |∈ S, v ∈ Ti

    Define ∆i as the star of vi in ∆, and let ∆PSi ⊂ ∆PS be its

    PS 6-split.

    Theorem

    Let s ∈ S12(∆PSi ). Let s be the restriction of |v |2s(v/|v |) to Ti .

    Then s is in S12(R−1i ∆

    PSi ) and

    s(vi) = s(vi), Dgi s(vi) = Dgi s(vi), Dhi s(vi) = Dhi s(vi).

  • Powell–Sabin splines Spherical Powell–Sabin splines Multiresolution analysis

    Spherical B-splines with control triangles

  • Powell–Sabin splines Spherical Powell–Sabin splines Multiresolution analysis

    Applications on a spherical domain

    Approximation of a mesh: consider the triangles of the originaltriangular mesh as control triangles of a PS spline.

  • Powell–Sabin splines Spherical Powell–Sabin splines Multiresolution analysis

    Applications on a spherical domain

    Compression by smoothing: Decimate a given triangular meshand approximate the decimated mesh.

    triangular mesh reduced mesh

    (40000 triangles) (5000 triangles)

  • Powell–Sabin splines Spherical Powell–Sabin splines Multiresolution analysis

    Applications on a spherical domain

    Compression by smoothing: Decimate a given triangular meshand approximate the decimated mesh.

    triangular mesh Powell–Sabin spline

    (40000 triangles) (5000 control triangles)

  • Powell–Sabin splines Spherical Powell–Sabin splines Multiresolution analysis

    Applications on a spherical domain

    triangular mesh decimated mesh spherical(40000 triangles) (5000 triangles) parameterization

    (5000 control triangles) PS spline surface

  • Powell–Sabin splines Spherical Powell–Sabin splines Multiresolution analysis

    Multiresolution analysis

  • Powell–Sabin splines Spherical Powell–Sabin splines Multiresolution analysis

    Multiresolution analysis (1989)

    Stéphane Mallat Yves Meyer

    A nested sequence of subspaces

    S0 ⊂ S1 ⊂ S2 ⊂ · · · ⊂ S` ⊂ · · ·

  • Powell–Sabin splines Spherical Powell–Sabin splines Multiresolution analysis

    Multiresolution analysis (1989)

    Stéphane Mallat Yves Meyer

    Complement spaces W`

    S`+1 = S` ⊕W`

  • Powell–Sabin splines Spherical Powell–Sabin splines Multiresolution analysis

    Multiresolution analysis (1989)

    Stéphane Mallat Yves Meyer

    A stable basis for the complement space W`

    W` = span{ψ`,i}

  • Powell–Sabin splines Spherical Powell–Sabin splines Multiresolution analysis

    Multiresolution analysis

    Refine the triangulation ∆ and its PS 6-split ∆PS.

    Vi

    Rki

    Vk

    Rjk

    Vj

    Rij

    Zijk

    Vi

    Rki

    Vk

    Rjk

    Vj

    Rij

    Zijk

    dyadic refinement triadic refinement

  • Powell–Sabin splines Spherical Powell–Sabin splines Multiresolution analysis

    Multiresolution analysis

    Refine the triangulation ∆ and its PS 6-split ∆PS.

    Vi

    Vki

    Vk

    Vjk

    Vj

    Vij

    Zijk

    Vi

    Rki

    Vk

    Rjk

    Vj

    Rij

    Vik

    Vki

    Vkj

    Vjk

    Vji

    Vij

    Vijk

    dyadic refinement triadic refinement

  • Powell–Sabin splines Spherical Powell–Sabin splines Multiresolution analysis

    Multiresolution analysis

    Refine the triangulation ∆ and its PS 6-split ∆PS.

    Vi

    Vki

    Vk

    Vjk

    Vj

    Vij

    Zijk

    Vi

    Rki

    Vk

    Rjk

    Vj

    Rij

    Vik

    Vki

    Vkj

    Vjk

    Vji

    Vij

    Vijk

    dyadic refinement triadic refinement

  • Powell–Sabin splines Spherical Powell–Sabin splines Multiresolution analysis

    Multiresolution analysis with√

    3-refinement

    ∆PS0 ⊂ ∆PS1 ⊂ · · · ⊂ ∆

    PS` ⊂ · · ·

    S12(∆PS0 ) ⊂ S

    12(∆

    PS1 ) ⊂ · · · ⊂ S

    12(∆

    PS` ) ⊂ · · ·

  • Powell–Sabin splines Spherical Powell–Sabin splines Multiresolution analysis

    Multiresolution analysis with√

    3-refinement

    ∆PS0 ⊂ ∆PS1 ⊂ · · · ⊂ ∆

    PS` ⊂ · · ·

    S12(∆PS0 ) ⊂ S

    12(∆

    PS1 ) ⊂ · · · ⊂ S

    12(∆

    PS` ) ⊂ · · ·

  • Powell–Sabin splines Spherical Powell–Sabin splines Multiresolution analysis

    Multiresolution analysis with√

    3-refinement

    ∆PS0 ⊂ ∆PS1 ⊂ · · · ⊂ ∆

    PS` ⊂ · · ·

    S12(∆PS0 ) ⊂ S

    12(∆

    PS1 ) ⊂ · · · ⊂ S

    12(∆

    PS` ) ⊂ · · ·

  • Powell–Sabin splines Spherical Powell–Sabin splines Multiresolution analysis

    Multiresolution analysis with√

    3-refinement

    ∆PS0 ⊂ ∆PS1 ⊂ · · · ⊂ ∆

    PS` ⊂ · · ·

    S12(∆PS0 ) ⊂ S

    12(∆

    PS1 ) ⊂ · · · ⊂ S

    12(∆

    PS` ) ⊂ · · ·

  • Powell–Sabin splines Spherical Powell–Sabin splines Multiresolution analysis

    Multiresolution analysis with√

    3-refinement

    S`+1 = S` ⊕W`

    Large triangles control S0Small triangles control W0Local edit

    Resolution level 0

  • Powell–Sabin splines Spherical Powell–Sabin splines Multiresolution analysis

    Multiresolution analysis with√

    3-refinement

    S`+1 = S` ⊕W`

    Large triangles control S0Small triangles control W0Local edit

    Resolution level 1

  • Powell–Sabin splines Spherical Powell–Sabin splines Multiresolution analysis

    Multiresolution analysis with√

    3-refinement

    S`+1 = S` ⊕W`

    Large triangles control S0Small triangles control W0Local edit

    Resolution level 1

  • Powell–Sabin splines Spherical Powell–Sabin splines Multiresolution analysis

    The hierarchical basis

    {Vi ∈ ∆`} ⊂ {Vi ∈ ∆`+1}

    ∆PS` ⊂ ∆PS`+1

    S` := S12(∆PS` ), S` ⊂ S`+1

    S2 = S0 ⊕W0 ⊕W1

    Largest triangles control S0Medium triangles control W0Smallest triangles control W1

    PSspline.mpgMedia File (video/mpeg)

  • Powell–Sabin splines Spherical Powell–Sabin splines Multiresolution analysis

    The hierarchical basis

    Basis functions: S` = span{φ`,k : k = 1, . . . ,3N`}

    s`(x , y) = φ`c` =N∑̀i=1

    3∑j=1

    Bij`(x , y)cij`

    φ`+1 = [φo`+1 φ

    n`+1],

    φo`+1 correspond to old reused vertices of level `φn`+1 correspond to the new vertices of level `+ 1

    The set of splines

    φ0 ∪m⋃

    `=1

    φn`

    forms a hierarchical basis for Sm.

  • Powell–Sabin splines Spherical Powell–Sabin splines Multiresolution analysis

    Wavelets via the lifting scheme

    φ` = φ`+1P`φ`+1 =

    [φo`+1 φ

    n`+1

    ][φ` ψ`

    ]= φ`+1

    [P` Q`

    ] (Wim Sweldens, 1994)Lifting

    ψ` = φn`+1 − φ`U`

    with U` the update matrix. We find a relation of the form[φ` ψ`

    ]= φ`+1

    [P`

    [0`I`

    ]− P`U`

    ]

  • Powell–Sabin splines Spherical Powell–Sabin splines Multiresolution analysis

    The update step

    Problems

    Semi-orthogonality⇒ U` not sparseFix U` sparse⇒ ψ` local supportWant stability⇒ need 1 vanishing moment for ψ`Remaining orthogonality conditions approximated by leastsquares

  • Powell–Sabin splines Spherical Powell–Sabin splines Multiresolution analysis

    The update step

    Problems

    Semi-orthogonality⇒ U` not sparseFix U` sparse⇒ ψ` local supportWant stability⇒ need 1 vanishing moment for ψ`Remaining orthogonality conditions approximated by leastsquares

  • Powell–Sabin splines Spherical Powell–Sabin splines Multiresolution analysis

    The update step

    Problems

    U` not sparse⇒ ψ` no local supportFix U` sparse⇒ ψ` local supportWant stability⇒ need 1 vanishing moment for ψ`Remaining orthogonality conditions approximated by leastsquares

  • Powell–Sabin splines Spherical Powell–Sabin splines Multiresolution analysis

    The update step

    Problems

    Want local support⇒ U` sparseFix U` sparse⇒ ψ` local supportWant stability⇒ need 1 vanishing moment for ψ`Remaining orthogonality conditions approximated by leastsquares

  • Powell–Sabin splines Spherical Powell–Sabin splines Multiresolution analysis

    The update step

    Problems

    Want local support⇒ U` sparseFix U` sparse⇒ ψ` local supportWant stability⇒ need 1 vanishing moment for ψ`Remaining orthogonality conditions approximated by leastsquares

  • Powell–Sabin splines Spherical Powell–Sabin splines Multiresolution analysis

    The update step

    Problems

    Want local support⇒ U` sparseOrthogonalize w.r.t. scaling functions in the update stencil

    Want stability⇒ need 1 vanishing moment for ψ`Remaining orthogonality conditions approximated by leastsquares

  • Powell–Sabin splines Spherical Powell–Sabin splines Multiresolution analysis

    The update step

    Problems

    Want local support⇒ U` sparseOrthogonalize w.r.t. scaling functions in the update stencil

    Want stability⇒ need 1 vanishing moment for ψ`Remaining orthogonality conditions approximated by leastsquares

  • Powell–Sabin splines Spherical Powell–Sabin splines Multiresolution analysis

    The update step

    Problems

    Want local support⇒ U` sparseOrthogonalize w.r.t. scaling functions in the update stencil

    i.e. φ̃` has to reproduce constants

    Remaining orthogonality conditions approximated by leastsquares

  • Powell–Sabin splines Spherical Powell–Sabin splines Multiresolution analysis

    The update step

    Problems

    Want local support⇒ U` sparseOrthogonalize w.r.t. scaling functions in the update stencil

    An extra linear constraint

    Remaining orthogonality conditions approximated by leastsquares

  • Powell–Sabin splines Spherical Powell–Sabin splines Multiresolution analysis

    The update step

    Problems

    Want local support⇒ U` sparseOrthogonalize w.r.t. scaling functions in the update stencil

    An extra linear constraint

    Remaining orthogonality conditions approximated by leastsquares

  • Powell–Sabin splines Spherical Powell–Sabin splines Multiresolution analysis

    Spherical Powell–Sabin spline wavelets

    3 wavelets per vertex

  • Powell–Sabin splines Spherical Powell–Sabin splines Multiresolution analysis

    Applications

    −→

    Spherical scattereddata

    Spherical PS spline surfacewith multiresolution structure

  • Powell–Sabin splines Spherical Powell–Sabin splines Multiresolution analysis

    Applications

    Compression

    Original 26%

  • Powell–Sabin splines Spherical Powell–Sabin splines Multiresolution analysis

    Applications

    Denoising

    With noise Denoised

  • Powell–Sabin splines Spherical Powell–Sabin splines Multiresolution analysis

    Applications

    Multiresolution editing

    Coarse level edit Fine level edit

  • Powell–Sabin splines Spherical Powell–Sabin splines Multiresolution analysis

    Applications

    (a) Coarse part of (c) (b) Coarse part of (d)

    (c) Original surface (d) Coarse level edit of (c)

  • Powell–Sabin splines Spherical Powell–Sabin splines Multiresolution analysis

    Some references

    P. Alfeld, M. Neamtu, and L. L. Schumaker. Bernstein–Bézierpolynomials on spheres and sphere-like surfaces. Comput. AidedGeom. Design, 13:333–349, 1996.

    P. Dierckx. On calculating normalized Powell–Sabin B-splines. Comput.Aided Geom. Design, 15(1), 61–78, 1997.

    J. Maes and A. Bultheel. Modeling sphere-like manifolds with sphericalPowell–Sabin B-splines. Comput. Aided Geom. Design, to appear.

    M. Neamtu and L. L. Schumaker. On the approximation order of splineson spherical triangulations. Adv. in Comp. Math., 21:3–20, 2004.

    W. Sweldens. The lifting scheme: A construction of second generationwavelets. SIAM J. Math. Anal., 29(2):511–546, 1997.

    Powell--Sabin splinesBernstein--BézierThe space of Powell--Sabin splinesB-splines with control triangles

    Spherical Powell--Sabin splinesSpherical spline spacesThe space of spherical Powell--Sabin splines

    Multiresolution analysisMultiresolution analysisWavelets via the lifting schemeThe update stepThe waveletsApplicationsReferences