Evolution of T-spline level sets for meshing non-uniformly sampled and incomplete data
read more
Citations
Numerical method for shape optimization using T-spline based isogeometric method
Parallel and adaptive surface reconstruction based on implicit PHT-splines
Restricted Trivariate Polycube Splines for Volumetric Data Modeling
Direct reconstruction of displaced subdivision surface from unorganized points
A survey on the local refinable splines
References
Bilateral filtering for gray and color images
Geodesic active contours
Level Set Methods and Dynamic Implicit Surfaces
Geodesic Active Contours
Regularization of Inverse Problems
Related Papers (5)
3D-reconstruction of complex geological interfaces from irregularly distributed and noisy point data
Frequently Asked Questions (24)
Q2. What contributions have the authors mentioned in the paper "Evolution of t-spline level sets for meshing non-uniformly sampled and incomplete data" ?
Given a large set of unorganized point sample data, the authors propose a new framework for computing a triangular mesh representing an approximating piecewise smooth surface. This framework is based on the combination of two types of surface representations: triangular meshes, and T-spline level sets, which are implicit surfaces defined by refinable spline functions allowing T-junctions. Firstly, the authors construct an implicit representation of a smooth ( C in their case ) surface, by using an evolution process of T-spline level sets, such that the implicit surface captures the topology and outline of the object to be reconstructed. Secondly, the authors project each data point to the initial mesh, and get a scalar displacement field. Finally, the authors present an additional evolution process, which combines data-driven velocities and featurepreserving bilateral filters, in order to reproduce sharp features. The authors also show that various shape constraints, such as distance field constraints, range constraints and volume constraints can be naturally added to their framework, which is helpful to obtain a desired reconstruction result, especially when the given data contains noise and inaccuracies.
Q3. What have the authors stated for future works in "Evolution of t-spline level sets for meshing non-uniformly sampled and incomplete data" ?
Possible future work includes comparing results by using different forms of functions for the implicit representation of the base surface, and exploiting other kinds of a priori knowledge about the geometric properties of the surface to be reconstructed.
Q4. What is the way to avoid cracks between adjacent triangles of a mesh?
In order to avoid cracks between adjacent triangles of a mesh, the interpolated normal is used [24] to displace the surface, B-spline surfaces are fitted [36] to the mesh before the displacement mapping, displaced subdivision surfaces [37] are suggested which is based on the butterfly subdivision scheme.
Q5. How can the base surface be reconstructed?
The authors have shown that, with the help of different shape constraints, even non-uniformly sampled and incomplete data can be handled by the implicitly defined base surface.
Q6. What is the method for generating a displaced surface?
Given a smooth base surface S0, a displaced surface S can be generated by a scalar field (a displacement map), which specifies the displacement values along the normal directions of S0.
Q7. What is the requirement for applying the Marching Triangulation method?
The requirement for applying the Marching Triangulation method is that the function value f(x) and the gradient ∇f(x) are available for any point x in the function domain.
Q8. What is the use of bilateral filters?
As a non-iterative scheme for edge-preserving smoothing, the bilateral filter is used in [20] and [32] to denoise a given surface.
Q9. What is the volume constraint for the evolution of surfaces?
By defining a smooth monotonic function V (τ) with respect to the time τ such that V (0) = V0 and V (τ) → V∞ as τ → ∞, the volume of the surface will continuously converge to the desired value V∞.
Q10. How can the horse split into several disconnected components in (a) without using the range constraint?
6. The narrow and thin legs of the horse can easily split into several disconnected components in (a) without using the range constraint.
Q11. how do the authors find the initial level set function f?
The initial level set function f can be found by approximating it to the signed distance field of a bounding sphere or a rough offset of the surface to be reconstructed [57].
Q12. What is the use of the bilateral filter?
In [4], the bilateral filter in used for data denoising such that the filtered data points can be later connected into a mesh structure.
Q13. What are the main constraints in the evolution process of T-spline level sets?
In particular, it is shown how to incorporate distance field constraints, range constraints and volume constraints such that the T-spline level sets evolution will be more robust and effective when dealing with non-uniformly sampled and incomplete data.
Q14. How can the authors handle the surface with only one component?
Since the triangulation of the initial surface with only one component is easily obtained, all the other surfaces can be handled subsequently.
Q15. What is the e(d) function used to determine the evolution step size?
∆τ is chosen as∆τ = min(1, { h|v(xj , τ) · n(xj , τ)| }0≤j≤N0) (14)where h is a user-defined constant to indicate the maximum allowed evolution step size for each sample point on the T-spline level set.
Q16. What is the way to reconstruct a smooth base surface?
Possible future work includes comparing results by using different forms of functions for the implicit representation of the base surface, and exploiting other kinds of a priori knowledge about the geometric properties of the surface to be reconstructed.
Q17. How can the cutting hole be recovered?
7. By using their method, the cutting hole can be filled by the T-spline level set representation, and the sharp edges can be recovered by the bilateral evolution of the mesh, as shown in (d).
Q18. What is the procedure for generating the mesh representation of S0?
4.1 Initial Mesh Generation through Marching TriangulationAfter the smooth base surface S0 is obtained, the authors use the Marching Triangulation [26] method to generate the mesh representation of S0.
Q19. What is the volume constraint for the evolution of T-spline level sets?
Now the evolution of T-spline level sets is transformed into a least-squares problem (21) subject to the linear volume constraint.
Q20. What is the purpose of the bilateral filter?
Recall that their bilateral evolution is to obtain a mesh that meets two goals:1. It provides a good fit to the point set (pk)k=1,2,...,n. 2. It recovers sharp features by conducting bilateral fil-ters.
Q21. which computer graphics laboratory is used for their experiments?
The data sets used for their experiments are courtesy of Stanford computer graphics laboratory and UCI (University of California, Irvine) computer graphics laboratory.
Q22. What is the last phase of the algorithm?
Since the data points are already well approximated by the displacement mesh, the last phase of their algorithm (recovering sharp features) is discarded.
Q23. what is the evolution of a T-spline level set?
(8)Combining (5), (6) and (8), the authors get the evolution equation of T-spline level sets under the vector field v,∂f(x, τ)∂τ = −v(x, τ) · ∇f(x, τ), x ∈ Γ (f). (9)In their method, the authors always start the evolution of Tspline level sets from an initial level set, which contains all data points inside.
Q24. What are the constraints used in the previous work?
The authors note that these constraints have been used in their previous work [18] in 2D for dual evolution of planar B-spline curves and T-spline level sets.