Mesh Deformation Based on Discrete Differential Geometry Reporter: Zhongping Ji 4.4.2007.

Slides:



Advertisements
Similar presentations
3D Geometry for Computer Graphics
Advertisements

3D reconstruction.
Chapter 9: Vector Differential Calculus Vector Functions of One Variable -- a vector, each component of which is a function of the same variable.
Active Contours, Level Sets, and Image Segmentation
Discrete Differential Geometry Planar Curves 2D/3D Shape Manipulation, 3D Printing March 13, 2013 Slides from Olga Sorkine, Eitan Grinspun.
Mean Value Coordinates for Closed Triangular Meshes
Discrete Exterior Calculus. More Complete Introduction See Chapter 7 “Discrete Differential Forms for Computational Modeling” in the SIGGRAPH 2006 Discrete.
Informationsteknologi Wednesday, November 7, 2007Computer Graphics - Class 51 Today’s class Geometric objects and transformations.
Chris Hall Aerospace and Ocean Engineering
Xianfeng Gu, Yaling Wang, Tony Chan, Paul Thompson, Shing-Tung Yau
Geometry (Many slides adapted from Octavia Camps and Amitabh Varshney)
INFORMATIK Differential Coordinates for Interactive Mesh Editing Yaron Lipman Olga Sorkine Daniel Cohen-Or David Levin Tel-Aviv University Christian Rössl.
Optimizing Content-Preserving Projections for Wide-Angle Images ACM SIGGRAPH 2009 Robert Carroll (University of California, Berkeley) Maneesh Agrawal (University.
A Sketch-Based Interface for Detail-Preserving Mesh Editing Andrew Nealen Olga Sorkine Marc Alexa Daniel Cohen-Or.
Chapter 5 Orthogonality
Spectral embedding Lecture 6 1 © Alexander & Michael Bronstein
6. One-Dimensional Continuous Groups 6.1 The Rotation Group SO(2) 6.2 The Generator of SO(2) 6.3 Irreducible Representations of SO(2) 6.4 Invariant Integration.
Computer Graphics Recitation 5.
Computer Graphics Recitation 2. 2 The plan today Learn about rotations in 2D and 3D. Representing rotations by quaternions.
1 Numerical geometry of non-rigid shapes Spectral Methods Tutorial. Spectral Methods Tutorial 6 © Maks Ovsjanikov tosca.cs.technion.ac.il/book Numerical.
3-D Geometry.
Spectral Embedding Alexander Bronstein, Michael Bronstein
Non-Euclidean Embedding
Previously Two view geometry: epipolar geometry Stereo vision: 3D reconstruction epipolar lines Baseline O O’ epipolar plane.
Orthogonality and Least Squares
1 Numerical geometry of non-rigid shapes Non-Euclidean Embedding Non-Euclidean Embedding Lecture 6 © Alexander & Michael Bronstein tosca.cs.technion.ac.il/book.
PHY 042: Electricity and Magnetism
Laplacian Surface Editing
Computer Graphics: Programming, Problem Solving, and Visual Communication Steve Cunningham California State University Stanislaus and Grinnell College.
Intrinsic Parameterization for Surface Meshes Mathieu Desbrun, Mark Meyer, Pierre Alliez CS598MJG Presented by Wei-Wen Feng 2004/10/5.
ORDINARY DIFFERENTIAL EQUATION (ODE) LAPLACE TRANSFORM.
CHAPTER FIVE Orthogonality Why orthogonal? Least square problem Accuracy of Numerical computation.
Computer Graphics Group Tobias Weyand Mesh-Based Inverse Kinematics Sumner et al 2005 presented by Tobias Weyand.
1 Mesh Parametrization and Its Applications 동의대학교 멀티미디어공학과 김형석 포항공과대학교 ( 이윤진, 이승용 )
Course 12 Calibration. 1.Introduction In theoretic discussions, we have assumed: Camera is located at the origin of coordinate system of scene.
Intro to 3D Models Angel Angel: Interactive Computer Graphics5E © Addison-Wesley
Finite Element Method.
视觉的三维运动理解 刘允才 上海交通大学 2002 年 11 月 16 日 Understanding 3D Motion from Images Yuncai Liu Shanghai Jiao Tong University November 16, 2002.
AN ORTHOGONAL PROJECTION
CSci 6971: Image Registration Lecture 3: Images and Transformations March 1, 2005 Prof. Charlene Tsai.
Shape Deformation Reporter: Zhang, Lei 5/30/2006.
Functions of Several Variables Copyright © Cengage Learning. All rights reserved.
Global Parametrization of Range Image Sets Nico Pietroni, Marco Tarini, Olga Sorkine, Denis Zorin.
Course 13 Curves and Surfaces. Course 13 Curves and Surface Surface Representation Representation Interpolation Approximation Surface Segmentation.
Computer Graphics Some slides courtesy of Pierre Alliez and Craig Gotsman Texture mapping and parameterization.
INFORMATIK Laplacian Surface Editing Olga Sorkine Daniel Cohen-Or Yaron Lipman Tel Aviv University Marc Alexa TU Darmstadt Christian Rössl Hans-Peter Seidel.
Elementary Linear Algebra Anton & Rorres, 9th Edition
USSC3002 Oscillations and Waves Lecture 11 Continuous Systems
Stable, Circulation- Preserving, Simplicial Fluids Sharif Elcott, Yiying Tong, Eva Kanso, Peter Schröder, and Mathieu Desbrun.
Andrew Nealen / Olga Sorkine / Mark Alexa / Daniel Cohen-Or SoHyeon Jeong 2007/03/02.
David Levin Tel-Aviv University Afrigraph 2009 Shape Preserving Deformation David Levin Tel-Aviv University Afrigraph 2009 Based on joint works with Yaron.
1 Chapter 2: Geometric Camera Models Objective: Formulate the geometrical relationships between image and scene measurements Scene: a 3-D function, g(x,y,z)
Comparing Two Motions Jehee Lee Seoul National University.
Determining 3D Structure and Motion of Man-made Objects from Corners.
CS559: Computer Graphics Lecture 9: 3D Transformation and Projection Li Zhang Spring 2010 Most slides borrowed from Yungyu ChuangYungyu Chuang.
Recent Progress in Mesh Parameterization Speaker : ZhangLei.
Basic Geometric Nonlinearities Chapter Five - APPENDIX.
CSE 554 Lecture 8: Alignment
Transformations.
Lecture 3 Jitendra Malik
Morphing and Shape Processing
Spectral Methods Tutorial 6 1 © Maks Ovsjanikov
Structure from motion Input: Output: (Tomasi and Kanade)
C H A P T E R 3 Vectors in 2-Space and 3-Space
2D Geometric Transformations
Y. Lipman D. Levin D. Cohen-Or
Reflections in Coordinate Plane
9.1 TRANSFORMAIONS.
Y. Lipman D. Levin D. Cohen-Or
Structure from motion Input: Output: (Tomasi and Kanade)
Presentation transcript:

Mesh Deformation Based on Discrete Differential Geometry Reporter: Zhongping Ji

Leonhard Euler

Papers (A) Linear Rotation-invariant Coordinates for Meshes [SIG 2005] Yaron Lipman Olga Sorkine David Levin Daniel Cohen-Or Tel Aviv University (B) Volume and Shape Preservation via Moving Frame Manipulation [TOG 2007] Yaron Lipman Daniel Cohen-Or Gal Ran David Levin Tel Aviv University

About authors  A Ph.D. student at The School of Mathematical Sciences, Tel-Aviv University. His supervisors are Prof. David Levin and Prof. Danny Cohen-Or. 

Mesh deformation

( A ) Rotation-Invariant  A rotation-invariant representation for triangular mesh based on two discrete forms;  The first form locally discribes the tangential component of the surface;  The second form locally discribes the normal component of the surface.

Classical differential geometry  The first and second fundamental forms;  Given two fundamental forms I and II, and they hold some compatibility conditions, Gauss-Codazzi-Mainardi equations;  Given an initial right-hand frame, then solve a first-order linear PDE.

Discrete Forms  Notations

First form Coefficients: Parameterization: First form: The length and the signed angle

Second form Coefficients: Second form: Height function of a vertex above the tangent plane

Lemma  Given the discrete form coefficients and the orientation bits at vertex i, the 1- ring neighborhood of i is defined up to a rigid transformation.  Fixing the position of vertex i, and fixing the direction of one edge emanating from vertex i and the normal, uniquely defines all the rest of the 1-ring vertex positions.

Proof

Discrete surface equations  Define a discrete frame at vertex i, Discrete surface equations: (1)

Theorem 1  The coefficients of the discrete surface equations can be expressed by the discrete forms.

Theorem 2  Given an initial discrete frame at an arbitrary vertex i 0, and given that the first and second discrete forms with the orientation bits are taken from an existing mesh, there exists a unique solution to the set of discrete surface equations.

Theorem 3 Given the discrete frame at each vertex, there exists a unique embedding of the vertices of the mesh in R 3 up to a translation, such that the resulting mesh has the given discrete forms and discrete frames.

Proof More constraints can be added to this equation. Geometry difference equations:

Mesh representation  The first discrete form is represented by the lengths of the projected edges and angles between adjacent projected edges.  The second discrete form represents the orthogonal component.  This sums to 3d i scalars, and the scalars are rigid motion invariant.

To reconstruct the mesh given its discrete form coefficients, an initial discrete frame and a position of one vertex in space. Construct the discrete surface equations; Add the initial discrete frame as an additional equation; Solve the augmented linear system in the least-squares sense to obtain the discrete frames; Construct the geometry difference equations; Add the given initial vertex position as an additional equation; Solve the augmented linear system in the least-squares sense to get the positions of the vertices.

Mesh editing  Two constraints and two sparse linear systems; Transformation constraints on the local frames and translation constraints on the vertex positions; Discrete surface equations and geometry difference equations.

Examples of basic editing operations

Results

Compatible constraints

Shape interpolation

Future work  Directly manipulate the local frames;  Improve the co-existence of constraints on the local frames and the vertex positions;  Use this representation for detail editing, such as detail enhancement.

( B ) Moving Frame  Isometric shape-preserving  Aiming at minimizing distortion of the shape.  Local volume preservation  Control the local volume by scaling the moving frames.

Motivation  Look for a deformation which maintains the first fundamental form intact (isometry), and minimizes changes in the second fundamental form.  Based on the Stokes’ theorem, using a carefully designed differential form which establishes a connetion between local volum and surface properties.

Shape-preserving  Isometric deformation Defining a rigid motion-invariant geometric distance(GD) between two isometric surfaces; Given a surface and a set of rotational constraints, look for an isometric deformation minimizing the GD to the original surface under the constraints.

Shape operator Orthonormal frame: Shape operator: Second fundamental form: Tangent plane:

Geometric distance Point to point: Surface to surface: meaning: measures to what extent the surface M and deformed surface are rigid motion of each other. Geometric deformation problem: Given a surface M, the goal is to deform it into another surface, subject to some prescribed constraints such that the geometric distance between them is minimal.

Solve the problem Dirichlet-type integral: An energy functional on the rotation’s field. The minimizer of the intergral, subject to constraints, is a harmonic function defined on a Lie group.

Shape-preserving method  Compute a rotation field R such that the enerty is minimized subject to constraints;  Apply rotations over the moving frames, and reconstruct the isometry f.

Parametrization of SO(3)  Dim(SO(n)) = n(n-1)/2  Two parametrization: orthogonal parametrization; conformal parametrization.

Orthogonal Parametrization

Solution

Conformal Parametrization

Solution

Piecewise-Linear Case (I)

Piecewise-Linear Case (II) Iterative scheme

An example

Volume preservation  Motivation, Strokes ’ theorem:  Designing a two-form: is a volumetric form and reflects the local volume underlying the point p on surface M.

Notations

Requirements: (1) (2) Illustration:

Derive two-form  Let (b 1,b 2,b 3 ) denote the standard basis in R 3,

Calculate scaling factor Scale the moving frame, scaling factor:

Calculate h(t) Defining the depth function Solve the ODE

Implementation  Select the set of Constraints: static and handle vertices;  Create and factorize the Laplace-Beltrami matrix of ROI;  Calculate the local depth field L of the mesh;  Calculate the rotation field R: M  SO(3) and apply to the moving frames at each vertex;  Reconstruct the mesh in a least-squares sense;  Calculate the scaling factor for volume correction;  Scale the moving frames and reconstruct the mesh.

Results Without volume correction

Results Without volume correction

Results With volume correction

Results Comparison

Results Comparison

Results

Irregular mesh

Relative change in volume

Future work  It cannot preserve simultaneously the surface area and volume of an object. An exciting avenue for future research is to investigate orher complete local surface descriptors, whose preservation yields a shape preservation of both surface and volume.

Thank you!