A signal processing approach to fair surface design

Gabriel Taubin

article1995SIGGRAPH2,243 citations

Proposes a linear-time signal processing framework that smooths large polyhedral surfaces of arbitrary topology via low-pass filtering while preventing shrinkage and enforcing geometric constraints.

Listen

Digital 3D models extracted from sources such as medical scans and range sensors often produce massive meshes containing hundreds of thousands to millions of vertices. Smoothing these large, arbitrary shapes without distorting their overall volume has historically presented a major computational bottleneck. Existing techniques rely on global optimization or solving large equation systems that scale quadratically with mesh size, making interactive editing unaffordable, while standard local smoothing methods severely shrink the geometry.

The article demonstrates a fast, signal-processing approach to smoothing polyhedral surfaces of arbitrary topology without causing shrinkage. It sets out to provide a scalable algorithm that integrates seamlessly into interactive geometric design workflows.

The proposed method adapts classical Fourier analysis to surfaces by treating vertex coordinates as discrete signals and framing smoothing as low-pass filtering. Rather than performing full frequency decompositions or solving massive linear systems, the approach uses a two-step iterative filter. It pairs a standard positive smoothing step with a compensating negative scaling step. The evaluation demonstrates this technique on synthetic geometries, medical models such as spine scans, and meshes integrated with subdivision algorithms.

Key findings show that the algorithm operates with linear time and memory complexity, requiring only basic sparse matrix multiplications rather than quadratic optimizations. Crucially, the alternating two-step filter removes high-frequency noise and unwanted undulations while preserving the model's original volume and major features. The algorithm also supports precise design constraints; users can lock specific vertices, control surface normals, and create smooth deformations by modifying local neighborhood definitions or solving only very small, localized linear systems.

These results demonstrate that smoothing very large meshes is practically feasible for interactive applications, drastically reducing processing costs and execution times. Interactive computer-aided design tools can adopt this filter to eliminate the trade-off between smoothing quality and geometric shrinkage, outperforming prior subdivision pipelines that introduced surface ripples or collapsed object boundaries.

Engineering and graphics teams should adopt this low-pass filtering formulation to replace Gaussian smoothing steps in subdivision frameworks and interactive mesh editors. Future development should explore expanding this signal-processing foundation to replace computationally heavy variational formulations in surface reconstruction and physics-based shape fitting.

Confidence in the approach is high for standard symmetric mesh structures, where stability and spectral bounds are mathematically guaranteed. However, users should exercise caution when introducing widespread non-symmetric neighborhood constraints, as these alter eigenvalue distributions and require validation to prevent unexpected deformations.

IBM fourier linear systems low-pass filtering Surface Signal Surfaces

Taubin (1995).pdf

Paper: The emerging field of signal processing on graphs: Extending high-dimensional data analysis to networks and other irregular domains, David I Shuman et al. (2012). Reviewing this signal processing framework on graphs provides essential background on generalizing harmonic analysis to irregular domains before reading about surface fairing.

Paper: Surface simplification using quadric error metrics, Michael Garland et al. (1997). This paper builds directly upon surface processing concepts by introducing quadric error metrics for high-fidelity mesh simplification.
Paper: Progressive meshes, Hugues Hoppe (1996). This work extends surface mesh manipulation techniques by introducing progressive representations and multi-resolution level-of-detail management.

Cover for A signal processing approach to fair surface design

Abstract

In this paper we describe a new tool for interactive free-form fair surface design. By generalizing classical discrete Fourier analysis to two-dimensional discrete surface signals – functions defined on polyhedral surfaces of arbitrary topology –, we reduce the problem of surface smoothing, or fairing, to low-pass filtering. We describe a very simple surface signal low-pass filter algorithm that applies to surfaces of arbitrary topology. As opposed to other existing optimization-based fairing methods, which are computationally more expensive, this is a linear time and space complexity algorithm. With this algorithm, fairing very large surfaces, such as those obtained from volumetric medical data, becomes affordable. By combining this algorithm with surface subdivision methods we obtain a very effective fair surface design technique. We then extend the analysis, and modify the algorithm accordingly, to accommodate different types of constraints. Some constraints can be imposed without any modification of the algorithm, while others require the solution of a small associated linear system of equations. In particular, vertex location constraints, vertex normal constraints, and surface normal discontinuities across curves embedded in the surface, can be imposed with this technique.

Topics

fourier linear systems low-pass filtering optimization optimization problems topology network topology

Keywords

Surface Signal Surfaces Design Algorithm Vertices Smoothing Fair Fairing Matrix

1 INTRODUCTION
2 THE SIGNAL PROCESSING APPROACH
2.1 CLOSED CURVE FAIRING
2.2 SURFACE SIGNAL FAIRING
3 SUBDIVISION
4 CONSTRAINTS
4.1 INTERPOLATORY CONSTRAINTS
4.2 SMOOTH INTERPOLATION
4.3 SMOOTH DEFORMATIONS
4.4 HIERARCHICAL CONSTRAINTS
4.5 TANGENT PLANE CONSTRAINTS
4.6 GENERAL LINEAR CONSTRAINTS
5 CONCLUSIONS
REFERENCES
APPENDIX

Knowls

Knowl 1 — Non-Shrinking Low-Pass Surface Smoothing Algorithm
algorithm
An algorithm to smooth (fair) a polyhedral surface mesh with $n$ vertices $V = (v_1, \dots, v_n)^t \in \mathbb{R}^{n \times 3}$ and arbitrary topology in linear time and space without shrinkage. It applies $N$ iterations consisting of two consecutive weighted averaging steps: a shrinking step with positive scale factor $\lambda$ followed by an un-shrinking step with negative scale factor $\mu$ .
Input: Initial vertex coordinates $V \in \mathbb{R}^{n \times 3}$ , neighborhood structure $\{i^\star\}_{i=1}^n$ , edge weights $\{w_{ij}\}$ , positive scale factor $\lambda > 0$ , negative scale factor $\mu < -\lambda$ , iteration count $N$
Output: Smoothed vertex coordinates $V^{(N)} \in \mathbb{R}^{n \times 3}$
$V^{(0)} \leftarrow V$
for step $s = 1$ to $N$ do
    for each vertex $i \in \{1, \dots, n\}$ do
         $\Delta v_i \leftarrow \sum_{j \in i^\star} w_{ij} (v_j^{(s-1)} - v_i^{(s-1)})$
         $v_i' \leftarrow v_i^{(s-1)} + \lambda \Delta v_i$
    for each vertex $i \in \{1, \dots, n\}$ do
         $\Delta v_i' \leftarrow \sum_{j \in i^\star} w_{ij} (v_j' - v_i')$
         $v_i^{(s)} \leftarrow v_i' + \mu \Delta v_i'$
return $V^{(N)}$
The algorithm exhibits $O(n)$ time complexity and $O(n)$ space complexity per iteration on sparse surface meshes. Typical parameter settings are $k_{PB} \approx 0.1$ , $\lambda = 0.6307$ , and $\mu = \frac{\lambda}{\lambda k_{PB} - 1} \approx -0.6732$ .
Knowl 2 — Transfer Function and Pass-Band Frequency for Non-Shrinking Smoothing
theoretical result
The non-shrinking surface smoothing algorithm implements a polynomial transfer function $f(k)$ evaluated on the discrete surface Laplacian matrix $K = I - W$ :

$f(k) = (1 - \lambda k)(1 - \mu k)$

where $k \in [0, 2]$ represents spatial frequency (the eigenvalues of $K$ ), $\lambda > 0$ is a positive scale factor, and $\mu$ is a negative scale factor satisfying $\mu < -\lambda < 0$ .

Because $f(0) = 1$ and $\lambda + \mu < 0$ , the filter possesses a strictly positive pass-band cutoff frequency $k_{PB}$ defined by:

$k_{PB} = \frac{1}{\lambda} + \frac{1}{\mu} > 0$

at which $f(k_{PB}) = 1$ .

After $N$ iterations, the cumulative transfer function is $f(k)^N$ . In the pass-band region $k \in [0, k_{PB}]$ , $f(k)^N \approx 1$ , preserving low-frequency geometric shape without overall volume shrinkage. In the stop-band region $k \in (k_{PB}, 2]$ , the response attenuates toward zero, provided $|f(k)| < 1$ for all $k \in (k_{PB}, 2]$ . A fast and stable filter with $k_{PB} < 1$ can be designed by choosing $\lambda$ such that $f(1) = -f(2)$ .
Knowl 3 — Discrete Surface Laplacian and Eigenvalue Spectrum on Polyhedral Meshes
definition
For a polyhedral surface with $n$ vertices $V = \{v_1, \dots, v_n\}$ , a discrete surface signal is a vector $x = (x_1, \dots, x_n)^t \in \mathbb{R}^n$ . The discrete Laplacian of $x$ at vertex $v_i$ is defined by:

$\Delta x_i = \sum_{j \in i^\star} w_{ij} (x_j - x_i)$

where $i^\star$ is the set of neighboring vertex indices of $v_i$ , and $w_{ij}$ are positive weights satisfying $\sum_{j \in i^\star} w_{ij} = 1$ for all $i$ .

In matrix notation, $\Delta x = -K x = -(I - W) x$ , where $W = (w_{ij})$ is the row-stochastic weight matrix ( $w_{ij} = 0$ for $j \notin i^\star$ ) and $I$ is the $n \times n$ identity matrix.

Common weight choices on first-order mesh neighborhoods include:
1. Uniform weights: $w_{ij} = \frac{1}{|i^\star|}$ .
2. Edge-function weights: $w_{ij} = \frac{\phi(v_i, v_j)}{\sum_{h \in i^\star} \phi(v_i, v_h)}$ , where $\phi(v_i, v_j) = \phi(v_j, v_i) > 0$ can be edge length powers $\|v_i - v_j\|^\alpha$ (typically $\alpha = -1$ ) or the sum of areas of faces sharing the edge $(v_i, v_j)$ .
When the neighborhood structure is symmetric and weights satisfy $W = E D$ with symmetric $E$ and diagonal $D$ , $W$ is similar to a symmetric matrix ( $D^{1/2} W D^{-1/2}$ is symmetric). Consequently, $K = I - W$ has $n$ real eigenvalues $0 \le k_1 \le k_2 \le \dots \le k_n \le 2$ with a complete set of linearly independent real unit-length right eigenvectors $u_1, \dots, u_n$ , which correspond to the discrete Fourier vibration modes of the surface.
Knowl 4 — Smooth Interpolatory Constraints for Surface Fairing
model/method
To enforce exact vertex position constraints during surface fairing while maintaining smooth surface transitions around constrained vertices, the constrained smoothed signal $x_C^N$ for $m$ target vertex values $(x_1, \dots, x_m)^t$ at vertices $v_1, \dots, v_m$ ( $m \ll n$ ) is defined as:

$x_C^N = x^N + F_{nm} F_{mm}^{-1} \begin{pmatrix} x_1 - x_1^N \\ \vdots \\ x_m - x_m^N \end{pmatrix}$

where:
- $x^N = F x = f(K)^N x$ is the unconstrained smoothed signal resulting from $N$ filter iterations on the original signal $x$ .
- $F = f(K)^N$ is the filter matrix, which is diagonally dominant and non-singular.
- $F_{nm} = [F \delta_1, \dots, F \delta_m]$ is the $n \times m$ matrix whose columns are smoothed discrete impulse signals $\delta_i$ (where $(\delta_i)_j = 1$ if $j=i$ and $0$ otherwise).
- $F_{mm}$ is the $m \times m$ principal submatrix formed by the first $m$ rows and columns of $F$ .
The computation is carried out in $O(m \cdot n)$ time and $O(n)$ space without forming the full $n \times n$ matrix $F$ :
1. For each $j \in \{1, \dots, m\}$ , apply the fairing algorithm to standard basis vector $\delta_j$ to obtain column $F \delta_j$ ; store its first $m$ elements in column $j$ of $F_{mm}$ and discard the remaining $n - m$ elements.
2. Solve the $m \times m$ linear system:
$F_{mm} \begin{pmatrix} y_1 \\ \vdots \\ y_m \end{pmatrix} = \begin{pmatrix} x_1 - x_1^N \\ \vdots \\ x_m - x_m^N \end{pmatrix}$
1. Define $y = (y_1, \dots, y_m, 0, \dots, 0)^t \in \mathbb{R}^n$ .
2. Apply the fairing algorithm to $y$ to compute $F y$ .
3. Return the constrained smoothed signal $x_C^N = x^N + F y$ .
Knowl 5 — General Linear Constraints in Surface Fairing via Coordinate Transformation
model/method
To fair a discrete surface signal $x \in \mathbb{R}^n$ subject to $m$ independent general linear constraints $C x_C^N = c$ , where $C$ is an $m \times n$ matrix of rank $m$ ( $m \le n$ ) and $c \in \mathbb{R}^m$ , the problem is transformed into a standard interpolatory constraint problem.

The constraint matrix $C$ and signal $x$ are partitioned as:

$C = \begin{bmatrix} C_{(1)} & C_{(2)} \end{bmatrix}, \quad x = \begin{pmatrix} x_{(1)} \\ x_{(2)} \end{pmatrix}$

where $C_{(1)}$ is an $m \times m$ non-singular submatrix (identified via Gaussian elimination with full pivoting) acting on $x_{(1)} \in \mathbb{R}^m$ , and $C_{(2)}$ is the $m \times (n-m)$ submatrix acting on $x_{(2)} \in \mathbb{R}^{n-m}$ .

A change of basis $x \mapsto y$ is defined by:

$\begin{pmatrix} x_{(1)} \\ x_{(2)} \end{pmatrix} = \begin{pmatrix} y_{(1)} - C_{(1)}^{-1} C_{(2)} y_{(2)} \\ y_{(2)} \end{pmatrix}$

Substituting this coordinate change into the linear constraint $C_{(1)} x_{(1)} + C_{(2)} x_{(2)} = c$ yields:

$C_{(1)} \left( y_{(1)} - C_{(1)}^{-1} C_{(2)} y_{(2)} \right) + C_{(2)} y_{(2)} = C_{(1)} y_{(1)} = c$

which simplifies to the standard interpolatory constraint on the transformed coordinates:

$y_{(1)} = C_{(1)}^{-1} c$

The transformed vector $y_{(1)}$ is constrained using the smooth interpolation algorithm, and the final surface signal is recovered via the inverse coordinate transformation.
Knowl 6 — Tangent Plane and Surface Normal Constraints via Discrete Laplacian
model/method
For a smooth $C^2$ continuous surface, the curvilinear integral of vertex displacements around a shrinking closed curve $\gamma$ enclosing vertex $v_i$ satisfies:

$\lim_{|\gamma| \to 0} \frac{1}{|\gamma|} \int_{v \in \gamma} (v - v_i) \, dl(v) = \bar{\kappa}(v_i) N_i$

where $|\gamma|$ is the arc length of $\gamma$ , $\bar{\kappa}(v_i)$ is the mean curvature at $v_i$ , and $N_i$ is the unit surface normal vector.

On a polyhedral mesh, the discrete Laplacian vector $\Delta v_i = \sum_{j \in i^\star} w_{ij}(v_j - v_i)$ provides a discrete approximation of this integral, implying that $\Delta v_i$ is collinear with the surface normal vector at $v_i$ .

To constrain the faired surface to have a prescribed normal direction $N_i$ at vertex $v_i$ after $N$ fairing iterations, two linearly independent vectors $S_i$ and $T_i$ spanning the tangent plane orthogonal to $N_i$ are chosen. The normal constraint is then formulated as two linear scalar constraints on the smoothed coordinates $V^N$ :

$S_i^t \Delta v_i^N = 0 \quad \text{and} \quad T_i^t \Delta v_i^N = 0$

where $\Delta v_i^N = \sum_{j \in i^\star} w_{ij} (v_j^N - v_i^N)$ . These constraints are enforced using the general linear constraint framework.
Knowl 7 — Hierarchical and Non-Symmetric Neighborhood Constraints
model/method
Modifying the neighborhood structure $\{i^\star\}_{i=1}^n$ of a polyhedral mesh allows enforcing boundary preservation and surface creases without altering the core smoothing algorithm:
1. Fixed Vertex Constraints: Setting the neighborhood of vertex $v_i$ to empty ( $i^\star = \emptyset$ ) causes $\Delta x_i = 0$ . Vertex $v_i$ remains stationary throughout smoothing while neighboring vertices $v_j$ (with $i \in j^\star$ ) continue to reference $v_i$ . This fixes $v_i$ but produces a sharp, non-smooth corner.
2. Hierarchical Neighborhoods: Each vertex $v_i$ is assigned an integer label $l_i$ . A vertex $v_j$ sharing an edge or face with $v_i$ is included in $i^\star$ if and only if $l_i \le l_j$ :
  - If $l_i < l_j$ , then $j \in i^\star$ but $i \notin j^\star$ , creating a directed dependency.
  - Vertices with identical labels ( $l_i = l_j$ ) have mutual symmetric neighbor relations.
Applications include:
- Setting $l_i = 1$ for mesh boundary vertices and $l_i = 0$ for interior vertices forces boundary curves to smooth as independent 1D curves while interior vertices smoothly follow the boundary.
- Setting $l_i = 1$ along an embedded curve of internal edges allows the surface to smooth along each side of the curve while allowing a sharp surface normal discontinuity (crease) across the curve.
- Setting $l_i = 2$ at isolated vertices along a crease curve holds those specific vertices stationary.
Knowl 8 — Fair Subdivision Surfaces via Non-Shrinking Filtering
model/method
In recursive subdivision surface schemes (such as Catmull-Clark quadrilateral subdivision or Loop triangular subdivision), standard algorithms perform a single Gaussian smoothing step ( $x' = x + \lambda \Delta x$ ) after each topological refinement step. A single Gaussian step fails to attenuate high-frequency shape components, allowing initial skeleton symmetries (such as hexagonal artifacts) to persist, whereas multiple Gaussian smoothing steps lead to severe mesh shrinkage.

By substituting the Gaussian smoothing step with $N$ iterations of non-shrinking smoothing using alternating positive ( $\lambda > 0$ ) and negative ( $\mu < -\lambda < 0$ ) scale factors (with passband frequency $k_{PB} \approx 0.1$ and $\lambda = 0.6307$ ):
1. High-frequency geometric noise and skeleton artifacts are eliminated without shrinking the overall volume.
2. The user gains design flexibility through continuous filter parameters ( $k_{PB}, \lambda, N$ ), replacing the fixed, parameter-free behavior of classical subdivision rules.
Knowl 9 — Interactive Free-Form Surface Deformations via Smooth Offset Fields
model/method
Interactive free-form surface deformation (direct manipulation by vertex pulling) is formulated using the smooth interpolatory constraint framework.

Given target displacement positions $\tilde{v}_1, \dots, \tilde{v}_m \in \mathbb{R}^3$ for $m$ selected handle vertices $v_1, \dots, v_m$ , the deformed surface $V_C^N$ is computed as:

$V_C^N = V^N + F_{nm} F_{mm}^{-1} \begin{pmatrix} \tilde{v}_1 - v_1^N \\ \vdots \\ \tilde{v}_m - v_m^N \end{pmatrix}$

where $V^N$ is the unconstrained faired mesh and $F = f(K)^{N_{def}}$ represents the smoothing operator applied to unit impulse basis vectors $\delta_1, \dots, \delta_m$ .

The spatial support and extent of the deformation are controlled by the iteration count $N_{def}$ used when smoothing the deformation basis vectors. Using $N_{def} \approx 5 N$ (where $N$ is the number of smoothing iterations applied to the underlying mesh) yields smooth, globally coordinated deformations across the surface.

Coverage note — None was omitted; all key theoretical and algorithmic contributions, including non-shrinking filtering, spectral properties, constraint formulations, subdivision integration, and deformations, are covered.

References

1.C.L. Bajaj and Ihm. Smoothing polyhedra using implicit algebraic splines. Computer Graphics, pages 79–88, July 1992. (Proceedings SIGGRAPH’92).
2.H. Baker. Building surfaces of evolution: The weaving wall. International Journal of Computer Vision, 3:51–71, 1989.
3.P. Borrel. Simple constrained deformations for geometric modeling and interactive design. ACM Transactions on Graphics, 13(2):137–155, April 1994.
4.E. Catmull and J. Clark. Recursively generated B-spline surfaces on arbitrary topological meshes. Computer Aided Design, 10:350–355, 1978.
5.G. Celniker and D. Gossard. Deformable curve and surface finite-elements for free-form shape design. Computer Graphics, pages 257–266, July 1991. (Proceedings SIGGRAPH’91).
6.T.D. DeRose, M. Lounsbery, and J. Warren. Multiresolution analysis for surfaces of arbitrary topological type. Technical Report 93-10-05, Department of Computer Science and Enginnering, University of Washington, Seattle, 1993.
7.M. Do Carmo. Differential Geometry of Curves and Surfaces. Prentice Hall, 1976.
8.D. Doo and M. Sabin. Behaviour of recursive division surfaces near extraordinary points. Computer Aided Design, 10:356–360, 1978.
9.G. Golub and C.F. Van Loan. Matrix Computations. John Hopkins University Press, 2nd. edition, 1989.
10.G. Greiner and H.P. Seidel. Modeling with triangular b-splines. IEEE Computer Graphics and Applications, 14(2):56–60, March 1994.
11.A. Guéziec and R. Hummel. The wrapper algorithm: Surface extraction and simplification. In IEEE Workshop on Biomedical Image Analysis, pages 204–213, Seattle, WA, June 24–25 1994.
12.M. Halstead, M. Kass, and T. DeRose. Efficient, fair interpolation using catmull-clark surface.Computer Graphics,pages 35–44,August 1993. (Proceedings SIGGRAPH’93).
13.H. Hoppe, T. DeRose, T. Duchamp, J. McDonald, and W. Stuetzle. Mesh optimization. Computer Graphics, pages 19–26, August 1993. (Proceedings SIGGRAPH’93).
14.W.M. Hsu, J.F. Hughes, and H. Kaufman. Direct manipulation of free-form deformations. Computer Graphics, pages 177–184, July 1992. (Proceedings SIGGRAPH’92).
15.A.D. Kalvin. Segmentation and Surface-Based Modeling of Objects in Three-Dimensional Biomedical Images. PhD thesis, New York University, New York, March 1991.
16.M. Kass, A. Witkin, and D. Terzopoulos. Snakes: active contour models. International Journal of Computer Vision, 1(4):321–331, 1988.
17.T. Lindeberg. Scale-space for discrete signals. IEEE Transactions on Pattern Analysis and Machine Intelligence, 12(3):234–254, March 1990.
18.C. Loop. Smooth subdivision surfaces based on triangles. Master’s thesis, Dept. of Mathematics, University of Utah, August 1987.
19.C. Loop. A G1 triangular spline surface of arbitrary topological type. Computer Aided Geometric Design, 11:303–330, 1994.
20.C. Loop. Smooth spline surfaces over irregular meshes. Computer Graphics, pages 303–310, July 1994. (Proceedings SIGGRAPH’94).
21.W. Lorenson and H. Cline. Marching cubes: A high resolution 3d surface construction algorithm. Computer Graphics, pages 163–169, July 1987. (Proceedings SIGGRAPH).
22.M. Lounsbery, S. Mann, and T. DeRose. Parametric surface interpolation. IEEE Computer Graphics and Applications, 12(5):45–52, September 1992.
23.J. Menon. Constructive shell representations for freeform surfaces and solids. IEEE Computer Graphics and Applications, 14(2):24–36, March 1994.
24.H.P. Moreton and C.H. Séquin. Functional optimization for fair surface design. Computer Graphics, pages 167–176, July 1992. (Proceedings SIGGRAPH’92).
25.J. Oliensis. Local reproducible smoothing without shrinkage. IEEE Transactions on Pattern Analysis and Machine Intelligence, 15(3):307–312, March 1993.
26.A. Pentland and S. Sclaroff. Closed-form solutions for physically based shape modeling and recognition. IEEE Transactions on Pattern Analysis and Machine Intelligence, 13(7):715–729, July 1991.
27.E. Seneta. Non-Negative Matrices, An Introduction to Theory and Applications. John Wiley & Sons, New York, 1973.
28.L.A. Shirman and C.H. Séquin. Local surface interpolation with bezier patches. Computer Aided Geometric Design, 4:279–295, 1987.
29.W.J. Shroeder, A. Zarge, and W.E. Lorensen. Decimation of triangle meshes. Computer Graphics, pages 65–70, 1992. (Proceedings SIGGRAPH’92).
30.R.S. Szeliski, D. Tonnesen, and D. Terzopoulos. Modeling surfaces of arbitrary topology with dynamic particles. In Proceedings, IEEE Conference on Computer Vision and Pattern Recognition, pages 82–87, New York, NY, June 15–17 1993.
31.G. Taubin. Curve and surface smoothing without shrinkage. Technical Report RC-19536, IBM Research, April 1994. (also in Proceedings ICCV’95).
32.G. Taubin. Estimating the tensor of curvature of a surface from a polyhedral approximation. Technical Report RC-19860, IBM Research, December 1994. (also in Proceedings ICCV’95).
33.G. Taubin, T. Zhang, and G. Golub. Optimal polyhedral surface smoothing as filter design. (in preparation).
34.D. Terzopoulos and K. Fleischer. Deformable models. The Visual Computer, 4:306–311, 1988.
35.G. Turk. Re-tiling polygonal surfaces. Computer Graphics, pages 55–64, July 1992. (Proceedings SIGGRAPH’92).
36.G. Turk and M. Levoy. Zippered polygon meshes from range data. Computer Graphics, pages 311–318, July 1994. (Proceedings SIGGRAPH’94).
37.W. Welch and A. Witkin. Variational surface modeling. Computer Graphics, pages 157–166, July 1992. (Proceedings SIGGRAPH’92).
38.W. Welch and A. Witkin. Free-form shape design using triangulated surfaces. Computer Graphics, pages 247–256, July 1994. (Proceedings SIGGRAPH’94).
39.A.P. Witkin. Scale-space filtering. In Proceedings, 8th. International Joint Conference on Artificial Intelligence (IJCAI), pages 1019–1022, Karlsruhe, Germany, August 1983.
40.C.T. Zahn and R.Z. Roskies. Fourier descriptors for plane closed curves. IEEE Transactions on Computers, 21(3):269–281, March 1972.

Citation

MLA

Taubin, G. “A Signal Processing Approach to Fair Surface Design”. Proceedings of the 22nd Annual Conference on Computer Graphics and Interactive Techniques - SIGGRAPH '95, 1995, pp. 351–58, https://doi.org/10.1145/218380.218473.

APA

Taubin, G. (1995). A signal processing approach to fair surface design. Proceedings of the 22nd Annual Conference on Computer Graphics and Interactive Techniques - SIGGRAPH '95, 351–358. https://doi.org/10.1145/218380.218473

Chicago

Taubin, G. 1995. “A Signal Processing Approach to Fair Surface Design”. Proceedings of the 22nd Annual Conference on Computer Graphics and Interactive Techniques - SIGGRAPH '95, 351–58. https://doi.org/10.1145/218380.218473.

Harvard

Taubin, G. (1995) “A signal processing approach to fair surface design”, Proceedings of the 22nd annual conference on Computer graphics and interactive techniques - SIGGRAPH '95. ACM Press, pp. 351–358. Available at: https://doi.org/10.1145/218380.218473.

Vancouver

1. Taubin G (1995) A signal processing approach to fair surface design. In: Proceedings of the 22nd annual conference on Computer graphics and interactive techniques - SIGGRAPH '95. ACM Press, pp 351–358

BibTeX

@inproceedings{Taubin_1995, series={SIGGRAPH ’95}, title={A signal processing approach to fair surface design}, url={http://dx.doi.org/10.1145/218380.218473}, DOI={10.1145/218380.218473}, booktitle={Proceedings of the 22nd annual conference on Computer graphics and interactive techniques  - SIGGRAPH ’95}, publisher={ACM Press}, author={Taubin, Gabriel}, year={1995}, pages={351–358}, collection={SIGGRAPH ’95} }

Metadata:Crossref

Access the Paper

This paper is available from its original source. Click below to access the PDF.

Open PDF

A signal processing approach to fair surface design

Abstract

Topics

Keywords

Table of Contents

Knowls

Knowl 1 — Non-Shrinking Low-Pass Surface Smoothing Algorithm

Knowl 2 — Transfer Function and Pass-Band Frequency for Non-Shrinking Smoothing

Knowl 3 — Discrete Surface Laplacian and Eigenvalue Spectrum on Polyhedral Meshes

Knowl 4 — Smooth Interpolatory Constraints for Surface Fairing

Knowl 5 — General Linear Constraints in Surface Fairing via Coordinate Transformation

Knowl 6 — Tangent Plane and Surface Normal Constraints via Discrete Laplacian

Knowl 7 — Hierarchical and Non-Symmetric Neighborhood Constraints

Knowl 8 — Fair Subdivision Surfaces via Non-Shrinking Filtering

Knowl 9 — Interactive Free-Form Surface Deformations via Smooth Offset Fields

References

Citation

BibTeX

Access the Paper