Best simple octagonal distances in digital geometry
The author shows that out of the class of neighbourhood sequences which have the same characteristic value there exists exactly one metric which has a simple functional form and incidentally satisfies the metricity conditions too. A revision of the available results on octagonal distances is given and a characterization for such simple octagonal distances is derived. The author introduces new error analyses involving these simple metrics. The error between the octagonal and the true euclidean distances is estimated in the asymptotic order by using a continuous approximation of the octagonal metric. Minimization of the absolute and the relative errors of the average of these simple distances with regard to the euclidean norm are carried out to identify the best approximate diginal distance in 2-D digital geometry. Four different simple metrics are recommended for practical use in digital approximation.
- The t-cost distance in digital geometry
- Parallel distance transforms on a linear array architecture
- Approximating the Euclidean distance using non-periodic neighbourhood sequences
- General neighborhood sequences in \(\mathbb Z^n\)
- scientific article; zbMATH DE number 4208959 (Why is no real title available?)
- Non-traditional 2D grids in combinatorial imaging -- advances and challenges
- Digital distance functions on three-dimensional grids
- Octagonal distances for digital pictures
- Distances based on neighbourhood sequences in non-standard three-dimensional grids
- Path-based distance functions in \(n\)-dimensional generalizations of the face- and body-centered cubic grids
This page was built for publication: Best simple octagonal distances in digital geometry
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q1185937)
