Abstract
Rigid motions in \(\mathbb {R}^2\) are fundamental operations in 2D image processing. They satisfy many properties: in particular, they are isometric and therefore bijective. Digitized rigid motions, however, lose these two properties. To investigate the lack of injectivity or surjectivity and more generally their local behavior, we extend the framework initially proposed by Nouvel and Rémila to the case of digitized rigid motions. Yet, for practical applications, the relevant information is not global bijectivity, which is seldom achieved, but bijectivity of the motion restricted to a given finite subset of \(\mathbb {Z}^2\). We propose two algorithms testing that condition. Finally, because rotation angles are rarely given with infinite precision, we propose a third algorithm providing optimal angle intervals that preserve this restricted bijectivity.
Similar content being viewed by others
Notes
Our implementation of the forward and backward algorithms can be downloaded from http://doi.org/10.5281/zenodo.248742.
In Thibault et al. [14], an algorithm was proposed for computing hinge angles \(h_\mathbf {t}^{<}(\mathbf {p},\theta )\) and \(h_\mathbf {t}^{>}(\mathbf {p},\theta )\)—for \(\mathbf {t}=\mathbf {0}\)—in a logarithmic time, which can be improved to constant time while considering half grid lines which bound the closest digitization cell, i.e., \(\mathscr {C}(U(\mathbf {p}))\). Notice that the algorithm also needs a modification in the while loop condition, such that \(k_{\max } - k_{\min } \le 1\), to avoid an infinite loop for some points, e.g., \(\mathbf {p}= (1,0)\).
References
Anglin, W.S.: Using Pythagorean triangles to approximate angles. Am. Math. Mon. 95(6), 540–541 (1988)
Fredriksson, K.: Rotation Invariant Template Matching. Ph.D. thesis, University of Helsinki (2001)
Galarza, A.I.R., Seade, J.: Introduction to Classical Geometries. Birkhäuser, Cambridge (2007)
Hunter, D.J.: Essentials of Discrete Mathematics, 2nd edn. Jones & Bartlett Learning, Burlington (2010)
Jacob, M.A., Andres, E.: On discrete rotations. In: 5th International Workshop on Discrete Geometry for Computer Imagery, pp. 161–174 (1995)
Nouvel, B., Rémila, E.: On colorations induced by discrete rotations. In: DGCI, Proceedings, Lecture Notes in Computer Science, vol. 2886, pp. 174–183. Springer (2003)
Nouvel, B., Rémila, E.: Characterization of bijective discretized rotations. In: IWCIA, Proceedings, Lecture Notes in Computer Science, vol. 3322, pp. 248–259. Springer (2004)
Nouvel, B., Rémila, E.: Configurations induced by discrete rotations: periodicity and quasi-periodicity properties. Discrete Appl. Math. 147(2–3), 325–343 (2005)
Nouvel, B., Rémila, E.: Incremental and transitive discrete rotations. In: IWCIA, Lecture Notes in Computer Science, vol. 4040, pp. 199–213. Springer (2006)
Pluta, K., Moroz, G., Kenmochi, Y., Romon, P.: Quadric arrangement in classifying rigid motions of a 3D digital image. In: CASC, Lecture Notes in Computer Science, vol. 9890, pp. 426–443 (2016)
Pluta, K., Romon, P., Kenmochi, Y., Passat, N.: Bijective rigid motions of the 2D Cartesian grid. In: DGCI, Lecture Notes in Computer Science, vol. 9647, pp. 359–371. Springer (2016)
Roussillon, T., Cœurjolly, D.: Characterization of bijective discretized rotations by Gaussian integers. Research Report. LIRIS UMR CNRS 5205 (2016). https://hal.archives-ouvertes.fr/hal-01259826
Thibault, Y.: Rotations in 2D and 3D discrete spaces. Ph.D. thesis, Université Paris-Est (2010)
Thibault, Y., Kenmochi, Y., Sugimoto, A.: Computing upper and lower bounds of rotation angles from digital images. Pattern Recogn. 42(8), 1708–1717 (2009)
Yilmaz, A., Javed, O., Shah, M.: Object tracking: a survey. ACM Comput. Surv. 38(4), 1–45 (2006)
Author information
Authors and Affiliations
Corresponding author
Appendices
Appendix 1: Neighborhood Motion Maps for \(\mathcal {G}_1^U\) (4-Neighborhood Case)
Neighborhood motion maps for \(\mathcal {G}_1^U\) are depicted as label maps, for \(\theta \in \left( 0,\frac{\pi }{6}\right) \) in Fig. 18; and \(\theta \in \left( \frac{\pi }{6},\frac{\pi }{4}\right) \) in Fig. 19. Some neighborhood motion maps are symmetric with respect to the origin, i.e., neighborhood motion map of the indices (0, 0)
Appendix 2: Neighborhood Motion Maps for \(\mathcal {G}^U_{2}\) (8-Neighborhood Case)
Neighborhood motion maps for \(\mathcal {G}^U_{2}\) are depicted as label maps, for \(\theta \in (0, \alpha _1)\) in Fig. 20; \(\theta \in (\alpha _1, \alpha _2)\) in Fig. 21; \(\theta \in (\alpha _2, \alpha _3)\) in Fig. 22; \(\theta \in (\alpha _3, \alpha _4)\) in Fig. 23; and \(\theta \in \big (\alpha _4,\frac{\pi }{4}\big )\) in Fig. 24. For more information about angles \(\alpha _n, n \in [1, 4]\) we refer the reader to Fig. 10. Some neighborhood motion maps are symmetric with respect to the origin, i.e., the neighborhood motion map of the indices (0, 0).
Rights and permissions
About this article
Cite this article
Pluta, K., Romon, P., Kenmochi, Y. et al. Bijective Digitized Rigid Motions on Subsets of the Plane. J Math Imaging Vis 59, 84–105 (2017). https://doi.org/10.1007/s10851-017-0706-8
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10851-017-0706-8