Computing handle and tunnel loops with knot linking
Introduction
Many applications need to identify features such as ‘handles’ and ‘tunnels’ induced by the embedding of a connected closed surface in three dimensions. Quantitatively, their numbers can be computed from Betti numbers, for which efficient algorithms are known [1]. However, these numbers cannot provide any qualitative information about these features. To this end, we define a class of loops on , called handle and tunnel loops, that help identifying these features in the shape bounded by . Intuitively, a loop is a handle if it spans a disk (surface) in the bounded space bordered by . If one cuts along such a loop and fills the boundary with that disk, one eliminates a handle. Similarly, a tunnel loop spans a disk (surface) in the unbounded space bordered by , and its removal eliminates a tunnel. Fig. 1 shows three handle and three tunnel loops on a CAD surface. In this paper, we provide a formal definition of handle and tunnel loops in terms of homology groups, and provide topological analyses that lead to the algorithms for their detection and generation. Our algorithm can be a basis for applications that require one to recognize features such as handles in a shape and tunnels in its complement, or to simplify a shape topologically by eliminating insignificant handles and tunnels [2], [3], [4], [5], [6].
Researchers have looked into the problem of computing nontrivial loops on surfaces with various conditions. Vegter and Yap [7] and Dey and Schipper [8] gave linear time algorithms to compute polygonal schema whose removal cuts the surface into a disk. Erickson and Har-Peled [9] showed that computing graphs of shortest length whose removal cut the surface into a disk is NP-hard. Verdière and Lazarus [10] gave an algorithm for computing a system of loops on a surface which is shortest among the homotopy class of a given system. Yin, Jin, and Gu used universal covering spaces [11] to compute shortest cycles in a homotopy class. Erickson and Whittlesey [12] gave a greedy algorithm to compute the shortest system of loops, among all systems of loops, relaxing the homotopy condition.
The above works were mainly concerned with computing a set of non trivial loops while optimizing some metric on the surface. Our goal is different. We seek to compute only specific loops that are handles and/or tunnels. One fundamental difference is that the aforementioned works do not take into account the embedding whereas handle and tunnel loops become meaningful only for embedded surfaces . Moreover, a loop may change its classification if the embedding changes.
We formalize the ideas of handle and tunnel loops and provide an existence proof for them. We argue that the notion of handle and tunnel loops loses its intuitive meaning if the surface has a knotted embedding. We define graph retractable surfaces that avoid these knotted embeddings. These are surfaces whose interior and exterior deformation retract to graphs called core graphs. We present algorithms to detect and generate handle and tunnel loops on such surfaces. Specifically, the main contributions of this paper are:
Definition and existence. We provide a formal definition of handle and tunnel loops and prove their existence.
Detection. We characterize handle and tunnel loops on graph retractable surfaces in terms of their linking with the core graphs. This leads to an algorithm for detecting handle and tunnel loops.
Generation. We show that there exists a special class of handle and tunnel loops that link minimally with the core graphs and present an algorithm to compute them.
Implementation. We present an implementation of our algorithm which incorporates the geometry of the surface more intimately. The results of our implementation show that the method is effective in practice.
Application. We apply our algorithm to the problem of computing ‘handle’ and ‘tunnel’ features in shapes which can further be used for topology simplification. Again, the results show the effectiveness of the method.
Section snippets
Preliminaries
We state some standard concepts from topology. For a more detailed introduction, the interested reader may consult Munkres [13] or Hatcher [14].
Let be any topological space. A singular -simplex is defined as a continuous map where is the standard -simplex which is the convex hull of . Each is a vector in and its th component is , where if and 0 otherwise. A -chain is a finite linear combination of singular -simplices. In this paper we assume the
Definition and existence
Let be a connected, closed (compact and without boundary), and orientable surface. The genus of is the maximum number of disjoint simple loops whose removal does not disconnect . Two closed, connected and orientable surfaces are homeomorphic if and only if they have the same genus. To make our argument simple, let sit inside a three sphere , which is the compactification of . Being embedded in , the surface has to be orientable. It separates into two parts. Given an
Graph retractable surface
Although Theorem 1 assures the existence of handle loops and tunnel loops on all connected closed surfaces in , they do not bear intuitive meaning of handles and tunnels in ‘knotted’ surfaces. Fig. 3 shows such a surface. It is obtained by thickening a trefoil knot, namely is the boundary of the product, , of a trefoil and a 2-disk . The red loop in Fig. 3(a) is obtained by projecting the trefoil knot on the surface . In contrary to the natural intuition this loop is not a tunnel
Graph complement basis through linking
Since we assume both and deformation retract to core graphs, we study the first homology groups of graphs and their complements in . Consider a knot 1. Since is homeomorphic to a circle , one can assign an orientation to it. It is known that and is the generator for . We also know that . Let be another knot in that is disjoint from . The union is a link. Consider a regular projection of link to a plane. Each
Minimally linked loops
In this section, we give more constructive statements about the handle and tunnel loops for a graph retractable surface. These statements relate the linking numbers of the loops with the core graphs to their characterization and existence. Theorem 3 characterizes the handle and tunnel loops in terms of linking numbers. Theorem 4 provides the existence of a special class of handle and tunnel loops which link minimally with the core graphs. This leads to an algorithm to compute a set of loops
Topological algorithms
In this section, we present algorithms to detect and generate handle and tunnel loops for a graph retractable surface.
Assume and are given. In a preprocessing step we compute a set of knots from and using their spanning trees as indicated in Section 5. Assuming that and have edges altogether, computation of s take time.
Detection: Let be any given loop with edges on . By Theorem 3, if links and not , it is a handle loop. If links and not ,
Incorporating geometry
In this section, we give an implementation of the algorithm, assuming that is presented as a piecewise linear surface. The algorithm described in Section 7 computes topologically correct handle and tunnel loops without considering any geometric information. Hence the computed loops may not be very good geometrically. The implementation presented in this section incorporates geometry into the algorithm, namely it computes two sets of loops with small size. Although it does not guarantee that
Experimental results
We implemented our algorithm in C++. Fig. 8 shows only the tunnel and the handle loops for some complicated models. As we can see, all detected loops are geometrically relevant. More examples are shown for feature detection and topology simplification in Fig. 11 and Fig. 12 which we discuss later.
In Table 1 we provide the times for our code on different models. The test machine has 2.66 GHz Intel Xeon CPU and 4 GB memory. We observe that curve skeleton computation dominates the computation
Application
In this section we apply handle and tunnel loops to compute actual handle or tunnel features for shapes which can further be used for removing insignificant topologies.
Feature detection. The basic idea is to sweep the handle or tunnel loops over the surface appropriately. To compute a tunnel feature, we start with a tunnel loop. We run Dijkstra’s shortest path algorithm for multiple sources, where the starting points are the vertices on the tunnel loop. At any generic step, Dijkstra’s algorithm
Conclusions and future work
In this work, we define and prove the existence of loops on surfaces which represent handles and tunnels for the shape bordered by the surface. We characterize these loops on graph retractable surfaces, using linking with core graphs. These characterizations lead to algorithms for detecting and generating handle and tunnel loops.
Several open questions arise as a result of this research. Our implementation does not guarantee that the computed tunnel or handle loops are the shortest in length.
Acknowledgments
We acknowledge that several models used in this paper are taken from AIM@SHAPE repository. We also acknowledge Cindy Grimm for the Knotty Cup and the Flower models.
References (21)
- et al.
An incremental algorithm for betti numbers of simplicial complexes on the 3-sphere
Comput Aided Geom Design
(1995) - et al.
Structure preserving CAD model repair
Comput Graphics Forum
(2005) - El-Sana J, Varshney A. Controlled simplification of genus for polygonal models. In: Proc. IEEE visualization’97. 1997....
- Guskov I, Wood Z. Topological noise removal. In: Proc. graphics interface 2001. 2001. p....
- et al.
Removing excess topology from isosurfaces
ACM Trans Graphics
(2004) - et al.
Topology repair of solid models using skeletons
IEEE Trans Visual Comput Graphics
(2007) - Vegter G, Yap CK. Computational complexity of combinatorial surfaces. In: Proc annu. sympos. comput. geom. 1990. p....
- et al.
A new technique to compute polygonal schema for 2-manifolds with application to null-homotopy detection
Discrete Comput Geom
(1995) - et al.
Optimally cutting a surface into a disk
Discrete Comput Geom
(2004) - et al.
Optimal system of loops on an orientable surface
Discrete Comput Geom
(2005)
Cited by (6)
Generation of Homotopy Classes for Unconstrained 3D Wire Routing from Characteristic Loops
2023, CAD Computer Aided DesignA novel technique for cohomology computations in engineering practice
2013, Computer Methods in Applied Mechanics and EngineeringFast linking numbers for topology verification of loopy structures
2021, ACM Transactions on GraphicsTopology preserving thinning of cell complexes
2014, IEEE Transactions on Image ProcessingChoking loops on surfaces
2013, IEEE Transactions on Visualization and Computer GraphicsEuclidean geodesic loops on high-genus surfaces applied to the morphometry of vestibular systems
2011, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)