80
Degree Reduction for NURBS Symbolic Computation on Curves Xianming Chen, Richard F. Riesenfeld, Elaine Cohen School of Computing University of Utah SMI’06

Degree Reduction for NURBS Symbolic Computation on Curves

  • Upload
    others

  • View
    7

  • Download
    0

Embed Size (px)

Citation preview

Page 1: Degree Reduction for NURBS Symbolic Computation on Curves

Degree Reduction for NURBS SymbolicComputation on Curves

Xianming Chen, Richard F. Riesenfeld, Elaine Cohen

School of ComputingUniversity of Utah

SMI’06

Page 2: Degree Reduction for NURBS Symbolic Computation on Curves

Background Derivatives of Rational Curves Plane Curves Space Curves Bisectors Conclusion

Outline

1 Background

2 Derivatives of Rational Curves

3 Symbolic Computation on B-Spline Plane CurvesCurvature Zero Set of a Rational Plane CurveCurvature Critical Set of a Plane CurveThe Evolute of a Plane Curve

4 Symbolic Computation on B-Spline Space Curves

5 Point/Curve and Curve/Curve BisectorsPolynomial Formulation of the Linear System for Solvingthe Curve/Curve BisectorRuled Point/Curve Bisectors

6 Conclusion

Page 3: Degree Reduction for NURBS Symbolic Computation on Curves

Background Derivatives of Rational Curves Plane Curves Space Curves Bisectors Conclusion

Outline

1 Background

2 Derivatives of Rational Curves

3 Symbolic Computation on B-Spline Plane CurvesCurvature Zero Set of a Rational Plane CurveCurvature Critical Set of a Plane CurveThe Evolute of a Plane Curve

4 Symbolic Computation on B-Spline Space Curves

5 Point/Curve and Curve/Curve BisectorsPolynomial Formulation of the Linear System for Solvingthe Curve/Curve BisectorRuled Point/Curve Bisectors

6 Conclusion

Page 4: Degree Reduction for NURBS Symbolic Computation on Curves

Background Derivatives of Rational Curves Plane Curves Space Curves Bisectors Conclusion

NURBS Symbolic Computation

1 closed algebraic operations on NURBS.2 Primitive Operation: sum/difference, multiplication/division,

differentiation, and composition (on polynomials)3 Derived Operation: all above on rationals, dot product,

cross product4 Problem: Fast raising degree when rational B-splines

involved.1 sum/difference is additive in terms of the final degree.2 differentiation doubles degree.

Page 5: Degree Reduction for NURBS Symbolic Computation on Curves

Background Derivatives of Rational Curves Plane Curves Space Curves Bisectors Conclusion

NURBS Symbolic Computation

1 closed algebraic operations on NURBS.2 Primitive Operation: sum/difference, multiplication/division,

differentiation, and composition (on polynomials)3 Derived Operation: all above on rationals, dot product,

cross product4 Problem: Fast raising degree when rational B-splines

involved.1 sum/difference is additive in terms of the final degree.2 differentiation doubles degree.

Page 6: Degree Reduction for NURBS Symbolic Computation on Curves

Background Derivatives of Rational Curves Plane Curves Space Curves Bisectors Conclusion

NURBS Symbolic Computation

1 closed algebraic operations on NURBS.2 Primitive Operation: sum/difference, multiplication/division,

differentiation, and composition (on polynomials)3 Derived Operation: all above on rationals, dot product,

cross product4 Problem: Fast raising degree when rational B-splines

involved.1 sum/difference is additive in terms of the final degree.2 differentiation doubles degree.

Page 7: Degree Reduction for NURBS Symbolic Computation on Curves

Background Derivatives of Rational Curves Plane Curves Space Curves Bisectors Conclusion

NURBS Symbolic Computation

1 closed algebraic operations on NURBS.2 Primitive Operation: sum/difference, multiplication/division,

differentiation, and composition (on polynomials)3 Derived Operation: all above on rationals, dot product,

cross product4 Problem: Fast raising degree when rational B-splines

involved.1 sum/difference is additive in terms of the final degree.2 differentiation doubles degree.

Page 8: Degree Reduction for NURBS Symbolic Computation on Curves

Background Derivatives of Rational Curves Plane Curves Space Curves Bisectors Conclusion

Outline

1 Background

2 Derivatives of Rational Curves

3 Symbolic Computation on B-Spline Plane CurvesCurvature Zero Set of a Rational Plane CurveCurvature Critical Set of a Plane CurveThe Evolute of a Plane Curve

4 Symbolic Computation on B-Spline Space Curves

5 Point/Curve and Curve/Curve BisectorsPolynomial Formulation of the Linear System for Solvingthe Curve/Curve BisectorRuled Point/Curve Bisectors

6 Conclusion

Page 9: Degree Reduction for NURBS Symbolic Computation on Curves

Background Derivatives of Rational Curves Plane Curves Space Curves Bisectors Conclusion

x ′x ′x ′,x ′′x ′′x ′′,x ′′′x ′′′x ′′′: a simple derivation

xxx = pppw , where p is a polynomial B-spline curve, and w is a

polynomial B-spline function. The derivatives are,

x ′x ′x ′ =p′p′p′w − pppw ′

w2 ,

x ′′x ′′x ′′ =p′′p′′p′′w − pppw ′′

w2 + (p′p′p′w − pppw ′)(1

w2 )′

x ′′′x ′′′x ′′′ =p′′′p′′′p′′′w − pppw ′′′

w2 + 2(p′′p′′p′′w − pppw ′′)(1

w2 )′ + (p′p′p′w − pppw ′)(1

w2 )′′+

p′′p′′p′′w ′ − p′p′p′w ′′

w2 .

Page 10: Degree Reduction for NURBS Symbolic Computation on Curves

Background Derivatives of Rational Curves Plane Curves Space Curves Bisectors Conclusion

Introducing D’s

D1 = p′p′p′w − pppw ′

D2 = p′′p′′p′′w − pppw ′′

D3 = p′′′p′′′p′′′w − pppw ′′′

D21 = p′′p′′p′′w ′ − p′p′p′w ′′

(1)

Page 11: Degree Reduction for NURBS Symbolic Computation on Curves

Background Derivatives of Rational Curves Plane Curves Space Curves Bisectors Conclusion

x ′x ′x ′,x ′′x ′′x ′′,x ′′′x ′′′x ′′′ in D’s

x ′x ′x ′ = 1w2 D1

x ′′x ′′x ′′ = 1w2 D2 − 2w ′

w3 D1

x ′′′x ′′′x ′′′= 1w2 D3 − 4w ′

w3 D2 + 6w ′2−2ww ′′

w4 D1 + 1w2 D21

(2)

Page 12: Degree Reduction for NURBS Symbolic Computation on Curves

Background Derivatives of Rational Curves Plane Curves Space Curves Bisectors Conclusion

D21

A Result from [?]

For a space curve xxx , D21, as defined in Eq. (1), is spanned byD1 and D2.

Page 13: Degree Reduction for NURBS Symbolic Computation on Curves

Background Derivatives of Rational Curves Plane Curves Space Curves Bisectors Conclusion

x ′x ′x ′,x ′′x ′′x ′′,x ′′′x ′′′x ′′′: a critical observation

By Eq. (2), the i-th (i = 0, 1, 2) order derivative of a rationalB-spline xxx is essentially the polynomial Di up to the same scalefactor 1

w2 , and some additive terms involving lower orderderivatives.

Therefore, for many rational curve inquiries, x ′x ′x ′,x ′′x ′′x ′′,x ′′′x ′′′x ′′′ can bereplaced by D1, D2, D3, respectively.

Page 14: Degree Reduction for NURBS Symbolic Computation on Curves

Background Derivatives of Rational Curves Plane Curves Space Curves Bisectors Conclusion

Degree Comparison: x ′x ′x ′,x ′′x ′′x ′′,x ′′′x ′′′x ′′′ vs. D1, D2, D3

x ′x ′x ′ 2d D1 2d-1x ′′x ′′x ′′ 4d D2 2d-2x ′′′x ′′′x ′′′ 8d D3 2d-3[x ′x ′x ′ x ′′x ′′x ′′] 6d [D1 D2] 4d-3[x ′x ′x ′ x ′′x ′′x ′′ x ′′′x ′′′x ′′′] 14d [D1 D2 D3] 6d-6

Assuming the degree of rational B-spline curve xxx is d

Page 15: Degree Reduction for NURBS Symbolic Computation on Curves

Background Derivatives of Rational Curves Plane Curves Space Curves Bisectors Conclusion

Outline

1 Background

2 Derivatives of Rational Curves

3 Symbolic Computation on B-Spline Plane CurvesCurvature Zero Set of a Rational Plane CurveCurvature Critical Set of a Plane CurveThe Evolute of a Plane Curve

4 Symbolic Computation on B-Spline Space Curves

5 Point/Curve and Curve/Curve BisectorsPolynomial Formulation of the Linear System for Solvingthe Curve/Curve BisectorRuled Point/Curve Bisectors

6 Conclusion

Page 16: Degree Reduction for NURBS Symbolic Computation on Curves

Background Derivatives of Rational Curves Plane Curves Space Curves Bisectors Conclusion

Curvature Zero Set of a Rational Plane Curve

κ =[x ′x ′x ′ x ′′x ′′x ′′]|x ′x ′x ′|3

f1 = [p′p′p′w − pppw ′ p′′p′′p′′w − pppw ′′]

= [p′p′p′ p′′p′′p′′]w2 − [ppp p′′p′′p′′]w ′w − [p′ pp′ pp′ p]ww ′′

f2 = [p′p′p′ p′′p′′p′′]w − [ppp p′′p′′p′′]w ′ − [p′ pp′ pp′ p]w ′′

1 κ = 0 iff f0 = [x ′x ′x ′ x ′′x ′′x ′′] = 02 f0 = 0 iff f1 = [D1 D2] = 03 Eliminate common w(6= 0),

f1 = 0 iff f2 = 04 Degf2 = 3d − 3, Degf0 = 6d .

Page 17: Degree Reduction for NURBS Symbolic Computation on Curves

Background Derivatives of Rational Curves Plane Curves Space Curves Bisectors Conclusion

Curvature Zero Set of a Rational Plane Curve

κ =[x ′x ′x ′ x ′′x ′′x ′′]|x ′x ′x ′|3

f1 = [p′p′p′w − pppw ′ p′′p′′p′′w − pppw ′′]

= [p′p′p′ p′′p′′p′′]w2 − [ppp p′′p′′p′′]w ′w − [p′ pp′ pp′ p]ww ′′

f2 = [p′p′p′ p′′p′′p′′]w − [ppp p′′p′′p′′]w ′ − [p′ pp′ pp′ p]w ′′

1 κ = 0 iff f0 = [x ′x ′x ′ x ′′x ′′x ′′] = 02 f0 = 0 iff f1 = [D1 D2] = 03 Eliminate common w(6= 0),

f1 = 0 iff f2 = 04 Degf2 = 3d − 3, Degf0 = 6d .

Page 18: Degree Reduction for NURBS Symbolic Computation on Curves

Background Derivatives of Rational Curves Plane Curves Space Curves Bisectors Conclusion

Curvature Zero Set of a Rational Plane Curve

κ =[x ′x ′x ′ x ′′x ′′x ′′]|x ′x ′x ′|3

f1 = [p′p′p′w − pppw ′ p′′p′′p′′w − pppw ′′]

= [p′p′p′ p′′p′′p′′]w2 − [ppp p′′p′′p′′]w ′w − [p′ pp′ pp′ p]ww ′′

f2 = [p′p′p′ p′′p′′p′′]w − [ppp p′′p′′p′′]w ′ − [p′ pp′ pp′ p]w ′′

1 κ = 0 iff f0 = [x ′x ′x ′ x ′′x ′′x ′′] = 02 f0 = 0 iff f1 = [D1 D2] = 03 Eliminate common w(6= 0),

f1 = 0 iff f2 = 04 Degf2 = 3d − 3, Degf0 = 6d .

Page 19: Degree Reduction for NURBS Symbolic Computation on Curves

Background Derivatives of Rational Curves Plane Curves Space Curves Bisectors Conclusion

Curvature Zero Set of a Rational Plane Curve

κ =[x ′x ′x ′ x ′′x ′′x ′′]|x ′x ′x ′|3

f1 = [p′p′p′w − pppw ′ p′′p′′p′′w − pppw ′′]

= [p′p′p′ p′′p′′p′′]w2 − [ppp p′′p′′p′′]w ′w − [p′ pp′ pp′ p]ww ′′

f2 = [p′p′p′ p′′p′′p′′]w − [ppp p′′p′′p′′]w ′ − [p′ pp′ pp′ p]w ′′

1 κ = 0 iff f0 = [x ′x ′x ′ x ′′x ′′x ′′] = 02 f0 = 0 iff f1 = [D1 D2] = 03 Eliminate common w(6= 0),

f1 = 0 iff f2 = 04 Degf2 = 3d − 3, Degf0 = 6d .

Page 20: Degree Reduction for NURBS Symbolic Computation on Curves

Background Derivatives of Rational Curves Plane Curves Space Curves Bisectors Conclusion

Curvature Critical Set of a Plane Curve

κ2 =[x ′x ′x ′x ′′x ′′x ′′]2

|x ′x ′x ′|6

κ′ =(x ′x ′x ′ ··· x ′x ′x ′)[x ′x ′x ′ x ′′′x ′′′x ′′′] − 3(x ′x ′x ′ ··· x ′′x ′′x ′′)[x ′x ′x ′ x ′′x ′′x ′′]

|x ′x ′x ′|5

=g0

|x ′x ′x ′|5

g2 = [D1 D3]|D1|2w + [D1 D21]|D1|2w+

2w ′[D1 D2]|D1|2 − 3(D1 ··· D2)[D1 D2]w

1 κ′ = 0 or κ = 0 iff (κ2)′ = 02 actually [?],

κ′ = 0 iff g0 = 0

3 if xxx rational,g0 = 0 iff g1 = 0

4 Degree comparison:poly. rational

κ2 12(d − 1) 48dg0 4d − 6 26dg1 9d − 6

Page 21: Degree Reduction for NURBS Symbolic Computation on Curves

Background Derivatives of Rational Curves Plane Curves Space Curves Bisectors Conclusion

Curvature Critical Set of a Plane Curve

κ2 =[x ′x ′x ′x ′′x ′′x ′′]2

|x ′x ′x ′|6

κ′ =(x ′x ′x ′ ··· x ′x ′x ′)[x ′x ′x ′ x ′′′x ′′′x ′′′] − 3(x ′x ′x ′ ··· x ′′x ′′x ′′)[x ′x ′x ′ x ′′x ′′x ′′]

|x ′x ′x ′|5

=g0

|x ′x ′x ′|5

g2 = [D1 D3]|D1|2w + [D1 D21]|D1|2w+

2w ′[D1 D2]|D1|2 − 3(D1 ··· D2)[D1 D2]w

1 κ′ = 0 or κ = 0 iff (κ2)′ = 02 actually [?],

κ′ = 0 iff g0 = 0

3 if xxx rational,g0 = 0 iff g1 = 0

4 Degree comparison:poly. rational

κ2 12(d − 1) 48dg0 4d − 6 26dg1 9d − 6

Page 22: Degree Reduction for NURBS Symbolic Computation on Curves

Background Derivatives of Rational Curves Plane Curves Space Curves Bisectors Conclusion

Curvature Critical Set of a Plane Curve

κ2 =[x ′x ′x ′x ′′x ′′x ′′]2

|x ′x ′x ′|6

κ′ =(x ′x ′x ′ ··· x ′x ′x ′)[x ′x ′x ′ x ′′′x ′′′x ′′′] − 3(x ′x ′x ′ ··· x ′′x ′′x ′′)[x ′x ′x ′ x ′′x ′′x ′′]

|x ′x ′x ′|5

=g0

|x ′x ′x ′|5

g2 = [D1 D3]|D1|2w + [D1 D21]|D1|2w+

2w ′[D1 D2]|D1|2 − 3(D1 ··· D2)[D1 D2]w

1 κ′ = 0 or κ = 0 iff (κ2)′ = 02 actually [?],

κ′ = 0 iff g0 = 0

3 if xxx rational,g0 = 0 iff g1 = 0

4 Degree comparison:poly. rational

κ2 12(d − 1) 48dg0 4d − 6 26dg1 9d − 6

Page 23: Degree Reduction for NURBS Symbolic Computation on Curves

Background Derivatives of Rational Curves Plane Curves Space Curves Bisectors Conclusion

Curvature Critical Set of a Plane Curve

κ2 =[x ′x ′x ′x ′′x ′′x ′′]2

|x ′x ′x ′|6

κ′ =(x ′x ′x ′ ··· x ′x ′x ′)[x ′x ′x ′ x ′′′x ′′′x ′′′] − 3(x ′x ′x ′ ··· x ′′x ′′x ′′)[x ′x ′x ′ x ′′x ′′x ′′]

|x ′x ′x ′|5

=g0

|x ′x ′x ′|5

g2 = [D1 D3]|D1|2w + [D1 D21]|D1|2w+

2w ′[D1 D2]|D1|2 − 3(D1 ··· D2)[D1 D2]w

1 κ′ = 0 or κ = 0 iff (κ2)′ = 02 actually [?],

κ′ = 0 iff g0 = 0

3 if xxx rational,g0 = 0 iff g1 = 0

4 Degree comparison:poly. rational

κ2 12(d − 1) 48dg0 4d − 6 26dg1 9d − 6

Page 24: Degree Reduction for NURBS Symbolic Computation on Curves

Background Derivatives of Rational Curves Plane Curves Space Curves Bisectors Conclusion

Curvature Critical Set of a Plane Curve

Example: Polynomial Cubic B-spline

Page 25: Degree Reduction for NURBS Symbolic Computation on Curves

Background Derivatives of Rational Curves Plane Curves Space Curves Bisectors Conclusion

Curvature Critical Set of a Plane Curve

Computing κ′ = 0: Squaring Approach

g1 = (κ2)′

1 degree = 242 # ctrl pnts = 1503 2 extra roots (κ = 0, but

κ′ 6= 0)

Page 26: Degree Reduction for NURBS Symbolic Computation on Curves

Background Derivatives of Rational Curves Plane Curves Space Curves Bisectors Conclusion

Curvature Critical Set of a Plane Curve

Computing κ′ = 0: Our Method

g2 =[x ′x ′x ′ x ′′′x ′′′x ′′′]|x ′x ′x ′|2 − 3(x ′x ′x ′ ··· x ′′x ′′x ′′)[x ′x ′x ′ x ′′x ′′x ′′]

1 degree = 62 # ctrl pnts = 243 exactly 6 roots where

κ′ = 0

Page 27: Degree Reduction for NURBS Symbolic Computation on Curves

Background Derivatives of Rational Curves Plane Curves Space Curves Bisectors Conclusion

Curvature Critical Set of a Plane Curve

Example: (Rational) Ellipse

Page 28: Degree Reduction for NURBS Symbolic Computation on Curves

Background Derivatives of Rational Curves Plane Curves Space Curves Bisectors Conclusion

Curvature Critical Set of a Plane Curve

Computing κ′ = 0: Squaring Approach

g1 = (κ2)′

1 degree = 962 # ctrl pnts = 388

Page 29: Degree Reduction for NURBS Symbolic Computation on Curves

Background Derivatives of Rational Curves Plane Curves Space Curves Bisectors Conclusion

Curvature Critical Set of a Plane Curve

Computing κ′ = 0: Our Method

g2 = Eq.(3.5) in the paper1 degree = 122 # ctrl pnts = 52

Page 30: Degree Reduction for NURBS Symbolic Computation on Curves

Background Derivatives of Rational Curves Plane Curves Space Curves Bisectors Conclusion

The Evolute of a Plane Curve

EEE(x) = xxx +1κ

NNN

EEE(x) = xxx +|x ′x ′x ′|3

[x ′x ′x ′ x ′′x ′′x ′′]x ′rx ′rx ′r|x ′rx ′rx ′r|

= xxx +|x ′x ′x ′|2

[x ′x ′x ′ x ′′x ′′x ′′]x ′rx ′rx ′r

EEE(x) =pppw

+|D1|2

[D1 D2]

D1r

w2

=[D1 D2]w ppp + |D1|2 D1r

[D1 D2]w2

1 evolute of a rational curve xxx2 a rational reformulation, with

deg = 13d3 using D’s, resulting deg =

6d − 3.

Page 31: Degree Reduction for NURBS Symbolic Computation on Curves

Background Derivatives of Rational Curves Plane Curves Space Curves Bisectors Conclusion

The Evolute of a Plane Curve

EEE(x) = xxx +1κ

NNN

EEE(x) = xxx +|x ′x ′x ′|3

[x ′x ′x ′ x ′′x ′′x ′′]x ′rx ′rx ′r|x ′rx ′rx ′r|

= xxx +|x ′x ′x ′|2

[x ′x ′x ′ x ′′x ′′x ′′]x ′rx ′rx ′r

EEE(x) =pppw

+|D1|2

[D1 D2]

D1r

w2

=[D1 D2]w ppp + |D1|2 D1r

[D1 D2]w2

1 evolute of a rational curve xxx2 a rational reformulation, with

deg = 13d3 using D’s, resulting deg =

6d − 3.

Page 32: Degree Reduction for NURBS Symbolic Computation on Curves

Background Derivatives of Rational Curves Plane Curves Space Curves Bisectors Conclusion

The Evolute of a Plane Curve

EEE(x) = xxx +1κ

NNN

EEE(x) = xxx +|x ′x ′x ′|3

[x ′x ′x ′ x ′′x ′′x ′′]x ′rx ′rx ′r|x ′rx ′rx ′r|

= xxx +|x ′x ′x ′|2

[x ′x ′x ′ x ′′x ′′x ′′]x ′rx ′rx ′r

EEE(x) =pppw

+|D1|2

[D1 D2]

D1r

w2

=[D1 D2]w ppp + |D1|2 D1r

[D1 D2]w2

1 evolute of a rational curve xxx2 a rational reformulation, with

deg = 13d3 using D’s, resulting deg =

6d − 3.

Page 33: Degree Reduction for NURBS Symbolic Computation on Curves

Background Derivatives of Rational Curves Plane Curves Space Curves Bisectors Conclusion

The Evolute of a Plane Curve

Example: Evolute of Ellipse usingEEE(x) = xxx + |x ′x ′x ′|2

[x ′x ′x ′ x ′′x ′′x ′′]x ′

rx ′rx ′r

1 degree = 262 # ctrl pnts = 108

Page 34: Degree Reduction for NURBS Symbolic Computation on Curves

Background Derivatives of Rational Curves Plane Curves Space Curves Bisectors Conclusion

The Evolute of a Plane Curve

Example: Evolute of Ellipse usingEEE(x) = xxx + |D1|2

[D1 D2]x ′

rx ′rx ′r

The same evolute1 degree = 92 # ctrl pnts = 40

Page 35: Degree Reduction for NURBS Symbolic Computation on Curves

Background Derivatives of Rational Curves Plane Curves Space Curves Bisectors Conclusion

Outline

1 Background

2 Derivatives of Rational Curves

3 Symbolic Computation on B-Spline Plane CurvesCurvature Zero Set of a Rational Plane CurveCurvature Critical Set of a Plane CurveThe Evolute of a Plane Curve

4 Symbolic Computation on B-Spline Space Curves

5 Point/Curve and Curve/Curve BisectorsPolynomial Formulation of the Linear System for Solvingthe Curve/Curve BisectorRuled Point/Curve Bisectors

6 Conclusion

Page 36: Degree Reduction for NURBS Symbolic Computation on Curves

Background Derivatives of Rational Curves Plane Curves Space Curves Bisectors Conclusion

τ = [x ′x ′x ′x ′′x ′′x ′′x ′′′x ′′′x ′′′]|x ′x ′x ′×x ′′x ′′x ′′|2 f = [D1 D2 D3]

DDDt = xxx + ζx ′x ′x ′ DDDDDDDDDt = ppp+λD1w

SSSb = xxx + ζ(x ′x ′x ′ × x ′′x ′′x ′′) SSSb = ppp+λ(D1×D2)w

SSSn = xxx + ζ(x ′x ′x ′ × x ′′x ′′x ′′) × x ′x ′x ′ SSSn = ppp+λ(D1×D2)×D1w

DDDDDDDDDr = xxx + ζ(τ TTT + κBBB) DDDDDDDDDr = xxx + λ(

[D1 D2 D3]|D1×D2|2

D1 + D1×D2|D1|2

)Unfortunately, the focal surface does not has a rationalrepresentation, in general.

Page 37: Degree Reduction for NURBS Symbolic Computation on Curves

Background Derivatives of Rational Curves Plane Curves Space Curves Bisectors Conclusion

τ = [x ′x ′x ′x ′′x ′′x ′′x ′′′x ′′′x ′′′]|x ′x ′x ′×x ′′x ′′x ′′|2 f = [D1 D2 D3]

DDDt = xxx + ζx ′x ′x ′ DDDDDDDDDt = ppp+λD1w

SSSb = xxx + ζ(x ′x ′x ′ × x ′′x ′′x ′′) SSSb = ppp+λ(D1×D2)w

SSSn = xxx + ζ(x ′x ′x ′ × x ′′x ′′x ′′) × x ′x ′x ′ SSSn = ppp+λ(D1×D2)×D1w

DDDDDDDDDr = xxx + ζ(τ TTT + κBBB) DDDDDDDDDr = xxx + λ(

[D1 D2 D3]|D1×D2|2

D1 + D1×D2|D1|2

)Unfortunately, the focal surface does not has a rationalrepresentation, in general.

Page 38: Degree Reduction for NURBS Symbolic Computation on Curves

Background Derivatives of Rational Curves Plane Curves Space Curves Bisectors Conclusion

τ = [x ′x ′x ′x ′′x ′′x ′′x ′′′x ′′′x ′′′]|x ′x ′x ′×x ′′x ′′x ′′|2 f = [D1 D2 D3]

DDDt = xxx + ζx ′x ′x ′ DDDDDDDDDt = ppp+λD1w

SSSb = xxx + ζ(x ′x ′x ′ × x ′′x ′′x ′′) SSSb = ppp+λ(D1×D2)w

SSSn = xxx + ζ(x ′x ′x ′ × x ′′x ′′x ′′) × x ′x ′x ′ SSSn = ppp+λ(D1×D2)×D1w

DDDDDDDDDr = xxx + ζ(τ TTT + κBBB) DDDDDDDDDr = xxx + λ(

[D1 D2 D3]|D1×D2|2

D1 + D1×D2|D1|2

)Unfortunately, the focal surface does not has a rationalrepresentation, in general.

Page 39: Degree Reduction for NURBS Symbolic Computation on Curves

Background Derivatives of Rational Curves Plane Curves Space Curves Bisectors Conclusion

τ = [x ′x ′x ′x ′′x ′′x ′′x ′′′x ′′′x ′′′]|x ′x ′x ′×x ′′x ′′x ′′|2 f = [D1 D2 D3]

DDDt = xxx + ζx ′x ′x ′ DDDDDDDDDt = ppp+λD1w

SSSb = xxx + ζ(x ′x ′x ′ × x ′′x ′′x ′′) SSSb = ppp+λ(D1×D2)w

SSSn = xxx + ζ(x ′x ′x ′ × x ′′x ′′x ′′) × x ′x ′x ′ SSSn = ppp+λ(D1×D2)×D1w

DDDDDDDDDr = xxx + ζ(τ TTT + κBBB) DDDDDDDDDr = xxx + λ(

[D1 D2 D3]|D1×D2|2

D1 + D1×D2|D1|2

)Unfortunately, the focal surface does not has a rationalrepresentation, in general.

Page 40: Degree Reduction for NURBS Symbolic Computation on Curves

Background Derivatives of Rational Curves Plane Curves Space Curves Bisectors Conclusion

τ = [x ′x ′x ′x ′′x ′′x ′′x ′′′x ′′′x ′′′]|x ′x ′x ′×x ′′x ′′x ′′|2 f = [D1 D2 D3]

DDDt = xxx + ζx ′x ′x ′ DDDDDDDDDt = ppp+λD1w

SSSb = xxx + ζ(x ′x ′x ′ × x ′′x ′′x ′′) SSSb = ppp+λ(D1×D2)w

SSSn = xxx + ζ(x ′x ′x ′ × x ′′x ′′x ′′) × x ′x ′x ′ SSSn = ppp+λ(D1×D2)×D1w

DDDDDDDDDr = xxx + ζ(τ TTT + κBBB) DDDDDDDDDr = xxx + λ(

[D1 D2 D3]|D1×D2|2

D1 + D1×D2|D1|2

)Unfortunately, the focal surface does not has a rationalrepresentation, in general.

Page 41: Degree Reduction for NURBS Symbolic Computation on Curves

Background Derivatives of Rational Curves Plane Curves Space Curves Bisectors Conclusion

τ = [x ′x ′x ′x ′′x ′′x ′′x ′′′x ′′′x ′′′]|x ′x ′x ′×x ′′x ′′x ′′|2 f = [D1 D2 D3]

DDDt = xxx + ζx ′x ′x ′ DDDDDDDDDt = ppp+λD1w

SSSb = xxx + ζ(x ′x ′x ′ × x ′′x ′′x ′′) SSSb = ppp+λ(D1×D2)w

SSSn = xxx + ζ(x ′x ′x ′ × x ′′x ′′x ′′) × x ′x ′x ′ SSSn = ppp+λ(D1×D2)×D1w

DDDDDDDDDr = xxx + ζ(τ TTT + κBBB) DDDDDDDDDr = xxx + λ(

[D1 D2 D3]|D1×D2|2

D1 + D1×D2|D1|2

)Unfortunately, the focal surface does not has a rationalrepresentation, in general.

Page 42: Degree Reduction for NURBS Symbolic Computation on Curves

Background Derivatives of Rational Curves Plane Curves Space Curves Bisectors Conclusion

Outline

1 Background

2 Derivatives of Rational Curves

3 Symbolic Computation on B-Spline Plane CurvesCurvature Zero Set of a Rational Plane CurveCurvature Critical Set of a Plane CurveThe Evolute of a Plane Curve

4 Symbolic Computation on B-Spline Space Curves

5 Point/Curve and Curve/Curve BisectorsPolynomial Formulation of the Linear System for Solvingthe Curve/Curve BisectorRuled Point/Curve Bisectors

6 Conclusion

Page 43: Degree Reduction for NURBS Symbolic Computation on Curves

Background Derivatives of Rational Curves Plane Curves Space Curves Bisectors Conclusion

Polynomial Formulation of the Linear System for Solving the Curve/Curve Bisector

Bisector as Locus of Intersecting Points of 3Planes

Adapted from [?], x ′x ′x ′

xxx ′

xxx − xxx

BBB =

xxx ··· x ′x ′x ′

xxx ··· xxx ′|xxx |2−xxx2

2

, (3)

Applying Cramer’s rule, the bisector is,

BBB =(D0, D1, D2)

D, (4)

where D is the determinant of the LHS matrix, and Di

(i = 0, 1, 2) is that of the LHS matrix when its i-th column isreplaced by the RHS column vector.

Page 44: Degree Reduction for NURBS Symbolic Computation on Curves

Background Derivatives of Rational Curves Plane Curves Space Curves Bisectors Conclusion

Polynomial Formulation of the Linear System for Solving the Curve/Curve Bisector

Calculate the Degree: Polynomial Case

Only focusing on degree in s of xxx(s) (BBB has the same degree ins of xxx(s) as that in s), Eq. (3) can be rewritten as, d − 1 d − 1 d − 1

0 0 0d d d

BBB =

2d − 10

2d

. (5)

1 D is polynomial of deg 2d − 12 Di is polynomial of deg 3d − 13 BBB = (Di

D ) is rational of deg 3d − 1

Page 45: Degree Reduction for NURBS Symbolic Computation on Curves

Background Derivatives of Rational Curves Plane Curves Space Curves Bisectors Conclusion

Polynomial Formulation of the Linear System for Solving the Curve/Curve Bisector

Calculate the Degree: Polynomial Case

Only focusing on degree in s of xxx(s) (BBB has the same degree ins of xxx(s) as that in s), Eq. (3) can be rewritten as, d − 1 d − 1 d − 1

0 0 0d d d

BBB =

2d − 10

2d

. (5)

1 D is polynomial of deg 2d − 12 Di is polynomial of deg 3d − 13 BBB = (Di

D ) is rational of deg 3d − 1

Page 46: Degree Reduction for NURBS Symbolic Computation on Curves

Background Derivatives of Rational Curves Plane Curves Space Curves Bisectors Conclusion

Polynomial Formulation of the Linear System for Solving the Curve/Curve Bisector

Calculate the Degree: Polynomial Case

Only focusing on degree in s of xxx(s) (BBB has the same degree ins of xxx(s) as that in s), Eq. (3) can be rewritten as, d − 1 d − 1 d − 1

0 0 0d d d

BBB =

2d − 10

2d

. (5)

1 D is polynomial of deg 2d − 12 Di is polynomial of deg 3d − 13 BBB = (Di

D ) is rational of deg 3d − 1

Page 47: Degree Reduction for NURBS Symbolic Computation on Curves

Background Derivatives of Rational Curves Plane Curves Space Curves Bisectors Conclusion

Polynomial Formulation of the Linear System for Solving the Curve/Curve Bisector

Calculate the Degree: Rational Case

1 D still has deg 2d − 1, but is now rational.2 By degree elevation and knot refinement, Di is a triple

product of (degree only) [2d − 1 0 2d ], thus is rational ofdeg 4d − 1.

3 BBB = (DiD ) is rational of deg 6d − 2

Page 48: Degree Reduction for NURBS Symbolic Computation on Curves

Background Derivatives of Rational Curves Plane Curves Space Curves Bisectors Conclusion

Polynomial Formulation of the Linear System for Solving the Curve/Curve Bisector

Calculate the Degree: Rational Case

1 D still has deg 2d − 1, but is now rational.2 By degree elevation and knot refinement, Di is a triple

product of (degree only) [2d − 1 0 2d ], thus is rational ofdeg 4d − 1.

3 BBB = (DiD ) is rational of deg 6d − 2

Page 49: Degree Reduction for NURBS Symbolic Computation on Curves

Background Derivatives of Rational Curves Plane Curves Space Curves Bisectors Conclusion

Polynomial Formulation of the Linear System for Solving the Curve/Curve Bisector

Calculate the Degree: Rational Case

1 D still has deg 2d − 1, but is now rational.2 By degree elevation and knot refinement, Di is a triple

product of (degree only) [2d − 1 0 2d ], thus is rational ofdeg 4d − 1.

3 BBB = (DiD ) is rational of deg 6d − 2

Page 50: Degree Reduction for NURBS Symbolic Computation on Curves

Background Derivatives of Rational Curves Plane Curves Space Curves Bisectors Conclusion

Polynomial Formulation of the Linear System for Solving the Curve/Curve Bisector

A Polynomial Reformulation for Rational Case

Observation

xxx ′ and xxx − xxx in Eq. (3) are directions, and thus can be replacedwith D1 and ppp − wxxx , respectively.

Consequently, Eq. (3) can be transformed into, D1

xxx ′

ppp − xxxw

BBB =

ppp ··· D1,

w xxx ··· xxx ′,ppp2−(xxxw)2

2 ,

, (6)

where,

BBB = wBBB. (7)

Page 51: Degree Reduction for NURBS Symbolic Computation on Curves

Background Derivatives of Rational Curves Plane Curves Space Curves Bisectors Conclusion

Polynomial Formulation of the Linear System for Solving the Curve/Curve Bisector

A Polynomial Reformulation for Rational Case

Observation

xxx ′ and xxx − xxx in Eq. (3) are directions, and thus can be replacedwith D1 and ppp − wxxx , respectively.

Consequently, Eq. (3) can be transformed into, D1

xxx ′

ppp − xxxw

BBB =

ppp ··· D1,

w xxx ··· xxx ′,ppp2−(xxxw)2

2 ,

, (6)

where,

BBB = wBBB. (7)

Page 52: Degree Reduction for NURBS Symbolic Computation on Curves

Background Derivatives of Rational Curves Plane Curves Space Curves Bisectors Conclusion

Polynomial Formulation of the Linear System for Solving the Curve/Curve Bisector

Bisector of Rational Curves Has Degree 4d − 1

Only focusing on degree in s, this is,2d − 1 2d − 1 2d − 10 0 0d d d

BBB =

3d − 1d

2d

. (8)

Therefore, we are back to the polynomial situation!1 D is polynomial of deg 3d − 12 Di is polynomial of deg 4d − 13 BBB = BBB

w = (D0,D1,D2)Dw is rational of deg 4d − 1 (in s).

Page 53: Degree Reduction for NURBS Symbolic Computation on Curves

Background Derivatives of Rational Curves Plane Curves Space Curves Bisectors Conclusion

Polynomial Formulation of the Linear System for Solving the Curve/Curve Bisector

Bisector of Rational Curves Has Degree 4d − 1

Only focusing on degree in s, this is,2d − 1 2d − 1 2d − 10 0 0d d d

BBB =

3d − 1d

2d

. (8)

Therefore, we are back to the polynomial situation!1 D is polynomial of deg 3d − 12 Di is polynomial of deg 4d − 13 BBB = BBB

w = (D0,D1,D2)Dw is rational of deg 4d − 1 (in s).

Page 54: Degree Reduction for NURBS Symbolic Computation on Curves

Background Derivatives of Rational Curves Plane Curves Space Curves Bisectors Conclusion

Polynomial Formulation of the Linear System for Solving the Curve/Curve Bisector

Bisector of Rational Curves Has Degree 4d − 1

Only focusing on degree in s, this is,2d − 1 2d − 1 2d − 10 0 0d d d

BBB =

3d − 1d

2d

. (8)

Therefore, we are back to the polynomial situation!1 D is polynomial of deg 3d − 12 Di is polynomial of deg 4d − 13 BBB = BBB

w = (D0,D1,D2)Dw is rational of deg 4d − 1 (in s).

Page 55: Degree Reduction for NURBS Symbolic Computation on Curves

Background Derivatives of Rational Curves Plane Curves Space Curves Bisectors Conclusion

Polynomial Formulation of the Linear System for Solving the Curve/Curve Bisector

Bisector of Rational Curves Has Degree 4d − 1

Only focusing on degree in s, this is,2d − 1 2d − 1 2d − 10 0 0d d d

BBB =

3d − 1d

2d

. (8)

Therefore, we are back to the polynomial situation!1 D is polynomial of deg 3d − 12 Di is polynomial of deg 4d − 13 BBB = BBB

w = (D0,D1,D2)Dw is rational of deg 4d − 1 (in s).

Page 56: Degree Reduction for NURBS Symbolic Computation on Curves

Background Derivatives of Rational Curves Plane Curves Space Curves Bisectors Conclusion

Polynomial Formulation of the Linear System for Solving the Curve/Curve Bisector

Bisector of Rational Curves Has Degree 4d − 1

Only focusing on degree in s, this is,2d − 1 2d − 1 2d − 10 0 0d d d

BBB =

3d − 1d

2d

. (8)

Therefore, we are back to the polynomial situation!1 D is polynomial of deg 3d − 12 Di is polynomial of deg 4d − 13 BBB = BBB

w = (D0,D1,D2)Dw is rational of deg 4d − 1 (in s).

Page 57: Degree Reduction for NURBS Symbolic Computation on Curves

Background Derivatives of Rational Curves Plane Curves Space Curves Bisectors Conclusion

Ruled Point/Curve Bisectors

Point/Curve Bisector as Locus of 2 IntersectingPlanes

(x ′x ′x ′

xxx −QQQ

)BBB =

(xxx ··· x ′x ′x ′

|xxx |2−QQQ2

2

), (9)

Page 58: Degree Reduction for NURBS Symbolic Computation on Curves

Background Derivatives of Rational Curves Plane Curves Space Curves Bisectors Conclusion

Ruled Point/Curve Bisectors

Solving the Under-determined Liner System: Adda 3rd Plane

The 3rd plane passes through the space point and has thecorresponding generator as its normal [?]. Together with theother 2 planes, it gives an explicit directrix DDD of the ruledbisector, x ′x ′x ′

x ′x ′x ′ × (xxx −QQQ)xxx −QQQ

DDD =

xxx ··· x ′x ′x ′

QQQ ··· (x ′x ′x ′ × (xxx −QQQ))|xxx |2−QQQ2

2

. (10)

Page 59: Degree Reduction for NURBS Symbolic Computation on Curves

Background Derivatives of Rational Curves Plane Curves Space Curves Bisectors Conclusion

Ruled Point/Curve Bisectors

Explicit Directrix Appr.: Point/Rational-CurveRuled Bisector Has Deg 17d

Focusing only on degrees in s, 2d 2d 2d3d 3d 3dd d d

DDD =

3d3d2d

. (11)

1 DDD is rational of deg (3d + 3d + 2d) + (2d + 3d + d) = 14d2 GGG = (x ′x ′x ′ × (xxx −QQQ)) is rational of deg 3d3 BBB = DDD + tGGG is rational of deg 17d

Page 60: Degree Reduction for NURBS Symbolic Computation on Curves

Background Derivatives of Rational Curves Plane Curves Space Curves Bisectors Conclusion

Ruled Point/Curve Bisectors

Explicit Directrix Appr.: Point/Rational-CurveRuled Bisector Has Deg 17d

Focusing only on degrees in s, 2d 2d 2d3d 3d 3dd d d

DDD =

3d3d2d

. (11)

1 DDD is rational of deg (3d + 3d + 2d) + (2d + 3d + d) = 14d2 GGG = (x ′x ′x ′ × (xxx −QQQ)) is rational of deg 3d3 BBB = DDD + tGGG is rational of deg 17d

Page 61: Degree Reduction for NURBS Symbolic Computation on Curves

Background Derivatives of Rational Curves Plane Curves Space Curves Bisectors Conclusion

Ruled Point/Curve Bisectors

Explicit Directrix Appr.: Point/Rational-CurveRuled Bisector Has Deg 17d

Focusing only on degrees in s, 2d 2d 2d3d 3d 3dd d d

DDD =

3d3d2d

. (11)

1 DDD is rational of deg (3d + 3d + 2d) + (2d + 3d + d) = 14d2 GGG = (x ′x ′x ′ × (xxx −QQQ)) is rational of deg 3d3 BBB = DDD + tGGG is rational of deg 17d

Page 62: Degree Reduction for NURBS Symbolic Computation on Curves

Background Derivatives of Rational Curves Plane Curves Space Curves Bisectors Conclusion

Ruled Point/Curve Bisectors

Explicit Directrix Appr.: Point/Rational-CurveRuled Bisector Has Deg 17d

Focusing only on degrees in s, 2d 2d 2d3d 3d 3dd d d

DDD =

3d3d2d

. (11)

1 DDD is rational of deg (3d + 3d + 2d) + (2d + 3d + d) = 14d2 GGG = (x ′x ′x ′ × (xxx −QQQ)) is rational of deg 3d3 BBB = DDD + tGGG is rational of deg 17d

Page 63: Degree Reduction for NURBS Symbolic Computation on Curves

Background Derivatives of Rational Curves Plane Curves Space Curves Bisectors Conclusion

Ruled Point/Curve Bisectors

Explicit Directrix Appro with PolynomialReformulation.: Point/Rational-Curve RuledBisector Has Deg 10d − 3

D1

D1 × (ppp − wQQQ)ppp − wQQQ

wDDD =

ppp ··· D1

wQQQ ··· (D1 × (ppp − wQQQ))ppp2−wQQQ2

2

,

(12)

and,focusing only on degree in s, this is, 2d − 1 2d − 1 2d − 13d − 1 3d − 1 3d − 1

d d d

wDDD =

3d − 14d − 1

2d

.

Page 64: Degree Reduction for NURBS Symbolic Computation on Curves

Background Derivatives of Rational Curves Plane Curves Space Curves Bisectors Conclusion

Ruled Point/Curve Bisectors

Explicit Directrix Appro with PolynomialReformulation.: Point/Rational-Curve RuledBisector Has Deg 10d − 3

D1

D1 × (ppp − wQQQ)ppp − wQQQ

wDDD =

ppp ··· D1

wQQQ ··· (D1 × (ppp − wQQQ))ppp2−wQQQ2

2

,

(12)

and,focusing only on degree in s, this is, 2d − 1 2d − 1 2d − 13d − 1 3d − 1 3d − 1

d d d

wDDD =

3d − 14d − 1

2d

.

Page 65: Degree Reduction for NURBS Symbolic Computation on Curves

Background Derivatives of Rational Curves Plane Curves Space Curves Bisectors Conclusion

Ruled Point/Curve Bisectors

Solving the Under-determined Liner System:Moving One Column to RHS

(xxx ′

xxx − QQQ

)BBB =

(xxx ··· x ′x ′x ′ − x ′x ′x ′[2]BBB[2]

|xxx |2−QQQ2

2 − (xxx [2] −QQQ[2])BBB[2].

),

where the notation aaa, for any vector aaa, denotes the vector aaa withthe last component discarded.

Page 66: Degree Reduction for NURBS Symbolic Computation on Curves

Background Derivatives of Rational Curves Plane Curves Space Curves Bisectors Conclusion

Ruled Point/Curve Bisectors

Direct Appr.: Point/Rational-Curve Ruled BisectorHas Deg 4d − 1

Focusing only on degrees in s, this is,(d − 1 d − 1

d d

)BBB =

(2d − 1

2d

). (13)

Page 67: Degree Reduction for NURBS Symbolic Computation on Curves

Background Derivatives of Rational Curves Plane Curves Space Curves Bisectors Conclusion

Ruled Point/Curve Bisectors

Example: Point/Helix-Like-Curve Bisector byExplicit Directrix Appr.

Evolute of Ellipse1 degree = 172 # ctrl pnts = 108

Page 68: Degree Reduction for NURBS Symbolic Computation on Curves

Background Derivatives of Rational Curves Plane Curves Space Curves Bisectors Conclusion

Ruled Point/Curve Bisectors

Example: Point/Helix-Like-Curve Bisector byDirect Appr.

The same evolute1 degree = 72 # ctrl pnts = 40

Page 69: Degree Reduction for NURBS Symbolic Computation on Curves

Background Derivatives of Rational Curves Plane Curves Space Curves Bisectors Conclusion

Ruled Point/Curve Bisectors

Example: Point/Ellipse Bisector by ExplicitDirectrix Appr.

Evolute of Ellipse1 degree = 172 # ctrl pnts = 108

Page 70: Degree Reduction for NURBS Symbolic Computation on Curves

Background Derivatives of Rational Curves Plane Curves Space Curves Bisectors Conclusion

Ruled Point/Curve Bisectors

Example: Point/Ellipse Bisector by Direct Appr.

The same evolute1 degree = 72 # ctrl pnts = 40

Page 71: Degree Reduction for NURBS Symbolic Computation on Curves

Background Derivatives of Rational Curves Plane Curves Space Curves Bisectors Conclusion

Ruled Point/Curve Bisectors

Example: Point/Ellipse Bisector Curve by ExplicitDirectrix Appr.

Evolute of Ellipse1 degree = 172 # ctrl pnts = 108

Page 72: Degree Reduction for NURBS Symbolic Computation on Curves

Background Derivatives of Rational Curves Plane Curves Space Curves Bisectors Conclusion

Ruled Point/Curve Bisectors

Example: Point/Ellipse Bisector Curve by DirectAppr.

The same evolute1 degree = 72 # ctrl pnts = 40

Page 73: Degree Reduction for NURBS Symbolic Computation on Curves

Background Derivatives of Rational Curves Plane Curves Space Curves Bisectors Conclusion

Outline

1 Background

2 Derivatives of Rational Curves

3 Symbolic Computation on B-Spline Plane CurvesCurvature Zero Set of a Rational Plane CurveCurvature Critical Set of a Plane CurveThe Evolute of a Plane Curve

4 Symbolic Computation on B-Spline Space Curves

5 Point/Curve and Curve/Curve BisectorsPolynomial Formulation of the Linear System for Solvingthe Curve/Curve BisectorRuled Point/Curve Bisectors

6 Conclusion

Page 74: Degree Reduction for NURBS Symbolic Computation on Curves

Background Derivatives of Rational Curves Plane Curves Space Curves Bisectors Conclusion

Conclusion1 presented several degree reduction strategies for NURBS

symbolic computation on curves, including1 eliminating higher degree terms resulting from irrelevant

lower order derivatives2 canceling common scalar factors3 polynomialization

2 an unaddressed numeric issue: the effect of discarding thedenominator to the final zero finding of the original rational

1 seems better for our examples.2 the opposite might be true if both the numerator and the

(discarded) denominator are evaluated too close to 0 atsome points.

Page 75: Degree Reduction for NURBS Symbolic Computation on Curves

Background Derivatives of Rational Curves Plane Curves Space Curves Bisectors Conclusion

Conclusion1 presented several degree reduction strategies for NURBS

symbolic computation on curves, including1 eliminating higher degree terms resulting from irrelevant

lower order derivatives2 canceling common scalar factors3 polynomialization

2 an unaddressed numeric issue: the effect of discarding thedenominator to the final zero finding of the original rational

1 seems better for our examples.2 the opposite might be true if both the numerator and the

(discarded) denominator are evaluated too close to 0 atsome points.

Page 76: Degree Reduction for NURBS Symbolic Computation on Curves

Background Derivatives of Rational Curves Plane Curves Space Curves Bisectors Conclusion

Conclusion1 presented several degree reduction strategies for NURBS

symbolic computation on curves, including1 eliminating higher degree terms resulting from irrelevant

lower order derivatives2 canceling common scalar factors3 polynomialization

2 an unaddressed numeric issue: the effect of discarding thedenominator to the final zero finding of the original rational

1 seems better for our examples.2 the opposite might be true if both the numerator and the

(discarded) denominator are evaluated too close to 0 atsome points.

Page 77: Degree Reduction for NURBS Symbolic Computation on Curves

Background Derivatives of Rational Curves Plane Curves Space Curves Bisectors Conclusion

Conclusion1 presented several degree reduction strategies for NURBS

symbolic computation on curves, including1 eliminating higher degree terms resulting from irrelevant

lower order derivatives2 canceling common scalar factors3 polynomialization

2 an unaddressed numeric issue: the effect of discarding thedenominator to the final zero finding of the original rational

1 seems better for our examples.2 the opposite might be true if both the numerator and the

(discarded) denominator are evaluated too close to 0 atsome points.

Page 78: Degree Reduction for NURBS Symbolic Computation on Curves

Background Derivatives of Rational Curves Plane Curves Space Curves Bisectors Conclusion

Conclusion1 presented several degree reduction strategies for NURBS

symbolic computation on curves, including1 eliminating higher degree terms resulting from irrelevant

lower order derivatives2 canceling common scalar factors3 polynomialization

2 an unaddressed numeric issue: the effect of discarding thedenominator to the final zero finding of the original rational

1 seems better for our examples.2 the opposite might be true if both the numerator and the

(discarded) denominator are evaluated too close to 0 atsome points.

Page 79: Degree Reduction for NURBS Symbolic Computation on Curves

Background Derivatives of Rational Curves Plane Curves Space Curves Bisectors Conclusion

Conclusion1 presented several degree reduction strategies for NURBS

symbolic computation on curves, including1 eliminating higher degree terms resulting from irrelevant

lower order derivatives2 canceling common scalar factors3 polynomialization

2 an unaddressed numeric issue: the effect of discarding thedenominator to the final zero finding of the original rational

1 seems better for our examples.2 the opposite might be true if both the numerator and the

(discarded) denominator are evaluated too close to 0 atsome points.

Page 80: Degree Reduction for NURBS Symbolic Computation on Curves

Background Derivatives of Rational Curves Plane Curves Space Curves Bisectors Conclusion

Conclusion1 presented several degree reduction strategies for NURBS

symbolic computation on curves, including1 eliminating higher degree terms resulting from irrelevant

lower order derivatives2 canceling common scalar factors3 polynomialization

2 an unaddressed numeric issue: the effect of discarding thedenominator to the final zero finding of the original rational

1 seems better for our examples.2 the opposite might be true if both the numerator and the

(discarded) denominator are evaluated too close to 0 atsome points.