1: \documentclass[11pt]{article}
2: \usepackage{epsfig}
3: \bibliographystyle{plain}
4: \textheight 22cm
5: \textwidth 17cm
6: \oddsidemargin 0mm
7: \evensidemargin 0mm
8: \headheight 0pt
9: \headsep 0pt
10: \topmargin 0mm
11: \parindent 1cm
12: \renewcommand{\textfraction}{.1}
13: \renewcommand{\topfraction}{.9}
14: \renewcommand{\bottomfraction}{.9}
15: \renewcommand{\baselinestretch}{1.35}
16: \newlength{\textbildsep}
17: \setlength{\textbildsep}{3.3ex}
18: \baselineskip=1.5\normalbaselineskip
19: \usepackage{amssymb}
20: \usepackage{amsmath}
21:
22:
23: \newtheorem{theorem}{Theorem}
24: \newtheorem{lemma}{Lemma}
25: \newtheorem{corollary}{Corollary}
26: \newtheorem{remark}{Remark}
27: \newtheorem{proposition}{Proposition}
28: \newtheorem{example}{Example}
29: \newtheorem{definition}{Definition}
30: \newtheorem{fact}{Fact}
31:
32: \input mssymb.tex
33: \def\E{\Bbb E}
34: \def\F{\Bbb F}
35: \def\R{\Bbb R}
36: \def\Z{\Bbb Z}
37: \def\Q{\Bbb Q}
38: \def\C{\Bbb C}
39: \def\G{\Bbb G}
40: \def\M{\Bbb M}
41: \def\N{\Bbb N}
42:
43: \begin{document}
44: \title{An explicit formula for the number of tunnels in digital objects}
45: \author{Valentin E. Brimkov\thanks{Mathematics Department, Buffalo State College, State University of New York,
46: 1300 Elmwood Ave., Buffalo, NY 14222, USA.
47: Email: brimkove@buffalostate.edu.}
48: \ \ \
49: Angelo Maimone$^{\dagger}$ \ \ \ Giorgio Nordo\thanks{Dipartimento di Matematica, Universit\`a di Messina, 98166 Messina, Italy.
50: E-mail: $\{$angelo.maimone,giorgio.nordo$\}$@unime.it.}}
51:
52: \date{}
53: \maketitle
54:
55: \begin{abstract}
56: An important concept in digital geometry for computer imagery is that of tunnel.
57: In this paper we obtain a formula for the number of tunnels
58: as a function of the number of the object
59: vertices, pixels, holes, connected components, and $2 \times 2$ grid squares.
60: It can be used to test for tunnel-freedom
61: a digital object, in particular a digital curve.
62:
63: {\bf Keywords:} {\em digital geometry, digital object, tunnel in a digital object}
64: \end{abstract}
65:
66: %---------------------
67: \section{Introduction}
68: %---------------------
69: \label{intro}
70: %---------------------
71:
72: An important concept in digital geometry for computer imagery is that of tunnel.
73: Intuitively, tunnels are locations in a digital object
74: (that is any finite set of pixels/voxels in 2D/3D) through which
75: a `discrete path' can penetrate.
76: Tunnels play an important role in rendering
77: pixelized/voxelized scenes by casting digital rays from the image to the scene
78: \cite{CohenKaufman}.
79: Thus it is useful to know if a digital object is tunnel-free or it has tunnels of certain type.
80: This is particularly interesting when dealing with digital curves or surfaces.
81: It may also be helpful to have an estimation for the number of tunnels (if any) in the
82: considered object, possibly as a function of other object characteristics.
83: Such kind of information may help better understand the structure of the object.
84: With this in mind, our objective was to obtain a formula that relates basic parameters of
85: a 2D digital object, such as the numbers of its pixels, vertices, holes, connected components, and tunnels.
86:
87: Results of this kind are of interest within several disciplines, such as digital geometry for computer imagery,
88: image analysis, computer graphics, and combinatorial geometry.
89: Let us mention, for instance, the famous
90: Euler's formula $v-e+f=2$ that relates the number of vertices, edges, and facets of a polytope,
91: and its applications to image analysis and digital geometry \cite{KLE2004}.
92: Other similar results are also available (see, e.g., Chapters 4 and 6 of \cite{KLE2004}).
93:
94: The main result of this paper is the formula $t=v-2(p+c-h)+b$, where $t$ is the number of tunnels,
95: $v$ the number of vertices, $p$ the number of pixels, $h$ the number of holes,
96: $c$ the number of connected components,
97: and $b$ the number of $2 \times 2$ grid squares (called 2-{\em blocks}) in a digital object.
98: This equality implies corollaries for the important cases
99: of simple closed digital curves, simple digital arcs, as well as for general digital curves.
100:
101: In the next section we introduce some basic notions and notations.
102: In Section 3 we present our main results, and we conclude in Section 4.
103:
104:
105:
106:
107: %---------------------
108: \section{Preliminaries}
109: %---------------------
110: \label{prelim}
111: %-------------
112:
113: \subsection{Basic Definitions}
114: \label{sec:def}
115: %--------------
116:
117: In this section we recall some basic notions of digital geometry following
118: \cite{KLE2004}. The reader is also referred to \cite{KON2001,BRI2002,BBN}.
119:
120: A regular orthogonal grid subdivides $\R^2$ into
121: unit squares called {\em pixels}, that are centered at the points of $\Z^2$.
122: Two pixels are called 0-{\em adjacent} if
123: they share a vertex, and 1-{\em adjacent} if they share an edge.
124: A digital object $S$ is a finite set of pixels.
125:
126: In the following definitions $k=0$ or 1.
127: A $k$-{\em path} in $S$ is a sequence of pixels from
128: $S$ such that every two consecutive pixels are $k$-adjacent.
129: Two pixels of a digital object $S$ are {\em $m$-connected} (in $S$)
130: iff there is a $k$-path in $S$ between them.
131: A digital object $S$ is {\em $k$-connected} iff
132: there is a $k$-path connecting any two pixels of $S$.
133: The maximal (by inclusion) $k$-connected subsets of a digital object $S$ are called
134: (connected) $k$-{\em components} of $S$.
135: Clearly, components are nonempty and distinct components are disjoint
136: with respect to $k$-adjacency.
137:
138:
139: Let $M$ be a subset of a digital object $S$.
140: If $S \setminus M$ is not $k$-connected then
141: the set $M$ is said to be $k$-{\em separating} in $S$.
142: Now let $M$ be a finite digital object that is $k$-separating in $\Z^2$.
143: The infinite 1-component of $\Z^2 \setminus M$ is called an {\em improper 1-hole} of $M$,
144: while the other (finite) 1-components of $S \setminus M$ are called {\em proper 1-holes} of $M$
145: (see Fig. \ref{connect0}).
146:
147: Let a digital object $M$ be $1$-separating but not $0$-separating
148: in a digital object $S$. Then $M$ is said to have $0$-{\em tunnels}
149: (see Fig. \ref{connect0}a, top-left).
150: For an object $M$ that is not separating in another object we can define a tunnel as
151: a point that is a common vertex of two and only two pixwls of $M$ (see Fig. \ref{connect0}a, top-right).
152: %----------------
153: \begin{figure}[t]
154: \begin{center}
155: \leavevmode
156: \epsfxsize=10cm
157: \epsffile{connect7.eps}
158: % \includegraphics[height=2.2cm]{fig4.ps}
159: \end{center}
160: \caption{a) {\em Top:} Digital curves with 0-tunnels.
161: {\em Bottom:} Tunnel-free digital curves. Each `0' features a hole. b) General digital curve with three holes.}
162: \label{connect0}
163: \end{figure}
164: %-----------
165: A digital object without any $0$-tunnels is called
166: {\em tunnel-free} (Fig. \ref{connect0}a, bottom).
167:
168:
169: %--------------------------
170: \subsection{Pixel Language}
171: %--------------------------
172: \label{language}
173: %---------------
174:
175: To facilitate the further description, below we present a simple graphical `language'
176: (similar to Venn diagrams in set theory), which we call the {\em pixel language}.
177: We will consider different configurations of pixels,
178: as in each there will be a `key' pixel, highlighted in grey,
179: whose neighborhood is studied.
180: Here are some key points of the pixel language
181: (see for illustration figures \ref{connect1},\ref{connect2}, and \ref{connect3}).
182:
183: \begin{itemize}
184: \item
185: Neighbors of the key pixel that {\em must} exist in the
186: considered configuration will be drawn with normal continuous lines.
187: \item
188: Pixels that {\em may} or {\em may not} exist in a configuration will be drawn with dashed lines.
189: Any subset of such pixels (in particular, no one or all of them)
190: may belong to the configuration or may be missing.
191: \item
192: Sometimes at least/at most/exactly one (or more) of these pixels will have to exist in a configuration.
193: In order to keep our pixel language simple, we will prefer to add relevant explanations in the text rather than
194: to introduce further special markings
195: (for other purposes it may be useful to do so).
196: \item
197: Grid positions that cannot contain pixels from the configuration will be marked by $\times$.
198: \item
199: Sometimes we will assign labels to pixels
200: and with their help analyze certain possibilities.
201: \item
202: Existing path in a digital object connecting pixels from the configuration will be marked
203: by a curve (as in some cases of figures 3 and 4).
204: \end{itemize}
205:
206:
207: \section{Tunnel formulas}
208: %---------------------
209: \subsection{Number of tunnels in arbitrary digital object}
210: \label{formula}
211: %------------
212: In this section we prove the following statement.
213:
214: \begin{theorem}
215: \label{Th1}
216: Consider a digital object $D \subset \Z^2$.
217: Let $p$ be the number of its pixels, $v$ the overal number of its pixels' vertices,
218: $h$ the number of its holes, $c$ the number of its connected components,
219: $b$ the number of its 2-blocks, and $t$ the number of its $0$-tunnels.
220: Then
221: \begin{equation}
222: \label{eq1}
223: t=v-2(p+c-h)+b.
224: \end{equation}
225: \end{theorem}
226: %-----------
227: \noindent {\bf Proof} \
228: We use induction on the number of pixels.
229: The statement is obviously true for a digital object consisting of a single pixel.
230: We have $p=1$, $v=4$, $c=1$, and $h = b = t = 0$, which values satisfy formula (\ref{eq1}).
231:
232: Assume that the statement is true for a digital object composed of $p$ pixels, where $p \geq 1$.
233: We will show that it is then true for any object composed by $p'=p+1$ pixels.
234: Consider such an object $D'$ and remove an arbitrary pixel $P$ from it.
235: Then $D=D'-\{P\}$ is a digital object of $p$ pixels to which the induction hypothesis applies.
236: Let the number of its vertices, components, holes, 2-blocks, and tunnels be $v,c,h,b$, and $t$, respectively.
237: Then $t=v-2(p+c-h)+b$. We will see how adding pixel $P$ to $D$ can influence this last equality.
238: We aim to show that
239: %------------------
240: \begin{equation}
241: \label{eq2}
242: t'=v'-2(p'+c'-h')+b',
243: \end{equation}
244: %------------------
245: where $p'=p+1, v', c', h', b'$, and $t'$ are the counts
246: of pixels, vertices, components, holes, 2-blocks, and tunnels of $D'$, respectively.
247: In doing so, we distinguish between 32 essentially different configurations
248: which we group into 10 cases, some of which involve
249: subcases.\footnote{In fact, there are a number of configurations that are not explicitly elaborated here,
250: but all of them fall within the considered 10 basic cases and admit analogous characterization. Their description
251: is differed to the full-length paper. Note also that any configuration containing dashed pixel(s) and/or path(s),
252: actually represents a number of different configurations, one for each possibility,
253: as all of them feature analogous properties.}
254: We analyze all of them with the help of illustrations using our pixel language.
255: For the sake of better readability of the paper,
256: we display the illustrations within three figures (number 2, 3, and 4), as labels of subfigures match the
257: numeration of the cases considered. Everywhere pixel $P$ is in dark grey.
258: We outline the main points of the proof. Details are diferred to the full length paper.
259:
260: Remember that throughout we have $p'=p+1$.
261: Upon adding $P$ to $D$, for the other parameters $v', c', h', b'$, and $t'$
262: we may have the following possibilities.
263:
264:
265: \medskip
266:
267: {\bf Case 1} \
268: {\em $c$, $h$, and $b$ do not change.}
269:
270: The only possible configurations under these conditions are displayed in Fig. \ref{connect1}.
271: We have $c'=c$, $h'=h$, and $b'=b$.
272:
273: In Case 1a we have $v'=v+2$ and $t'=t$.
274: In Case 1b, $v'=v+3$ and $t'=t+1$.
275: In Case 1c, $v'=v+1$ and $t'=t-1$.
276: In Case 1d, $v'=v$ and $t'=t-2$.
277:
278:
279: \medskip
280:
281: {\bf Case 2} \
282: {\em $h$ and $b$ do not change, while $c$ increases by 1.}
283:
284: Obviously, adding $P$ to $D$ can increase the number of components of $D'=D \cup \{ P \}$
285: only if $P$ is disjoint from $D$.
286: Then $c$ will increase by 1, while $h$, $b$, as well as $t$ will not change
287: (see Fig. \ref{connect1} (2)).
288:
289: We have $v'=v+4$, $c'=c+1$, $h'=h$, $b'=b$, and $t'=t$.
290:
291: \medskip
292:
293: {\bf Case 3} \
294: {\em $h$ and $b$ do not change, while $c$ decreases by 1,2, or 3.}
295:
296: The possible configurations are displayed in Fig. \ref{connect1}.
297: We have $h'=h$ and $b'=b$.
298:
299: In Case 3a we have $c'=c-1$, which is possible for two configurations displayed in Fig. \ref{connect1} (3a: top, middle, bottom).
300: For the first configuration we have $v'=v$ and $t'=t$, while for the second and the third we have $v'=v+2$ and $t'=t+2$,
301: respectively.
302: In Case 3b we have $c'=c-2$. The possible configuration is the one in Fig. \ref{connect1} (3b).
303: We have $v'=v+1$, $t'=t+3$.
304: In Case 3c we have $c'=c-3$ (Fig. \ref{connect1} (3c)).
305: We have $v'=v$, $t'=t+4$.
306:
307: %----------------
308: \begin{figure}[t]
309: \begin{center}
310: \leavevmode
311: \epsfxsize=8cm
312: \epsffile{connect1.eps}
313: % \includegraphics[height=2.2cm]{fig4.ps}
314: \end{center}
315: \caption{Illustrations to the proof of Theorem \ref{Th1}. Cases 1, 2, and 3.}
316: \label{connect1}
317: \end{figure}
318: %-----------
319:
320:
321: \medskip
322:
323:
324: {\bf Case 4} \
325: {\em $b$ and $c$ do not change, while $h$ decreases by 1.}
326:
327: The only possible configuration is displayed in Fig. \ref{connect2} (4).
328: We have $h'=h+1$, $c'=c$, $b'=b$, $v'=v$, $t'=t-4$.
329:
330: \medskip
331:
332: {\bf Case 5} \
333: {\em $c$ and $b$ do not change, while $h$ inecreases by 1,2, or 3.}
334:
335: The possible configurations are displayed in Fig. \ref{connect2}.
336: We have $c'=c$ and $b'=b$.
337:
338: In Case 5a we have $h'=h+1$, for which there are three possible configurations displayed in Fig. \ref{connect2} (5a: top, middle, bottom). For the first configuration, $v'=v$ and $t'=t$. For the second, $v'=v+1$ and $t'=t+1$. For the third, $v'=v$ and $t'=t$
339: (note that adding $P$ creates a new tunnel but closes an existing one).
340:
341:
342: In Case 5b we have $h'=h+2$, which is possible for two configurations displayed in Fig. \ref{connect2} (5b: top, bottom).
343: For the first one, $v'=v$ and $t'=t+2$, while for the second $v'=v+1$ and $t'=t+3$.
344:
345: In Case 5c we have $h'=h+3$, which may occur in two configurations (Fig. \ref{connect2} (5c)).
346: For both we have $v'=v$ and $t'=t+4$.
347: Note that in the first case three new holes may appear in a hole-free component,
348: while in the second an existing hole is partitioned into four holes.
349:
350: \medskip
351:
352: %----------------
353: \begin{figure}[t]
354: \begin{center}
355: \leavevmode
356: \epsfxsize=9cm
357: \epsffile{connect2.eps}
358: % \includegraphics[height=2.2cm]{fig4.ps}
359: \end{center}
360: \caption{Illustrations to the proof of Theorem \ref{Th1}. Cases 4, 5, 6, and 7.}
361: \label{connect2}
362: \end{figure}
363: %-----------
364:
365:
366: {\bf Case 6} \
367: {\em $c$ and $h$ do not change, while $b$ increases by 1 or 2.}
368:
369: In the former case there are two possible configurations displayed in Fig. \ref{connect2} (6a,b).
370: We have $c'=c$ and $h'=h$.
371: In Case 6a we have $b'=b+1$, $v'=v$, and $t'=t-1$ (Fig. \ref{connect2} (6a)).
372: In Case 6b we have $b'=b+1$, $v'=v+1$, and $t'=t$ (Fig. \ref{connect2} (6b)).
373:
374: In the latter case, we have the configuration displayed in Fig. \ref{connect2} (6c).
375: We have $b'=b+2$, $c'=c$, $h'=h$, $v'=v$, and $t'=t$.
376:
377: \medskip
378:
379:
380: {\bf Case 7} \
381: {\em $c$ does not change, while $h$ and $b$ increase by 1.}
382:
383: The only possible configuration is displayed in Fig. \ref{connect2} (7).
384: We have $v'=v$, $h'=h+1$, $b'=b+1$, $c'=c$, and $t'=t+1$.
385:
386:
387: \medskip
388:
389:
390:
391:
392: {\bf Case 8} \
393: {\em $c$ does not change, $h$ decreases by 1, and $b$ increases by 1,2,3, or 4.}
394:
395: The possible configurations are displayed in Fig. \ref{connect3}.
396: We have $c'=c$ and $h'=h-1$.
397:
398: In Case 8a we have $b'=b+1$, $v'=v$, and $t'=t-3$ (Fig. \ref{connect3} (8a)).
399: In Case 8b we have $b'=b+2$, $v'=v$, and $t'=t-2$ (Fig. \ref{connect3} (8b)).
400: In Case 8c we have $b'=b+3$, $v'=v$, and $t'=t-1$ (Fig. \ref{connect3} (8c)).
401: In Case 8d we have $b'=b+4$, $v'=v$, and $t'=t$ (Fig. \ref{connect3} (8d)).
402:
403: \medskip
404:
405: {\bf Case 9} \
406: {\em $h$ does not change, while $b$ increases by 1 and $c$ decreases by 1.}
407:
408: The only possible configuration is displayed in Fig. \ref{connect3} (9).
409: Note that the pixel marked by 1 is not connected to
410: the component to which the new pixel $P$ belongs.
411: We have $h'=h$, $b'=b+1$, $c'=c-1$, $v'=v$, and $t'=t+1$.
412:
413: \medskip
414:
415: {\bf Case 10} \
416: {\em $b$ does not change,
417: while $h$ increases by 1 and $c$ decreases by 1 (Case 10a),
418: or $h$ increases by 1 and $c$ decreases by 2 (Case 10b),
419: or $h$ increases by 2 and $c$ decreases by 1 (Case 10c).}
420:
421: The two possible configurations for Case 10a are displayed in Fig. \ref{connect3} (10a).
422: For the one on the top we have $v'=v$ and $t'=t+2$, while for the other (bottom) we have $v'=v+1$ and $t'=t+3$.
423: Case 10b features two possible configurations displayed in Fig. \ref{connect3} (10b).
424: For both we have $v'=v$, $t'=t+4$.
425: In Case 10c we have $v'=v$, $t'=t+4$ (Fig. \ref{connect3} (10c)).
426:
427: Note that in all figures numbered pixels belong to a component that is not connected to the pixel $P$.
428: %----------------
429: \begin{figure}[t]
430: \begin{center}
431: \leavevmode
432: \epsfxsize=9cm
433: \epsffile{connect3.eps}
434: % \includegraphics[height=2.2cm]{fig4.ps}
435: \end{center}
436: \caption{Illustrations to the proof of Theorem \ref{Th1}. Cases 8, 9, and 10.}
437: \label{connect3}
438: \end{figure}
439: %-----------
440:
441: \smallskip
442:
443: Having all possible cases determined, simple substitutions in (\ref{eq2}) for the respective values of
444: $v'$, $t'$, $h'$, $b'$, and $c'$ show that this last equality holds in all cases.
445:
446: \smallskip
447:
448: It is easy to realize that the considered cases are the only possible (up to certain symmetries).
449: Simple reasoning reveals that adding a pixel to $D$ can neither result in
450: decreasing $b$, nor in
451: increasing both $h$ and $c$, nor in
452: increasing both $b$ and $c$, nor in
453: decreasing $h$ and changing $c$.
454: This completes the proof of the theorem.
455: $\Box$
456: %-----
457: \begin{corollary}
458: Let $M$ be a tunnel-free digital object. Then $v-2(p+c-h)+b=0$.
459: \end{corollary}
460: %-------------
461:
462: %--------------------------
463: \subsection{Tunnels in curves}
464: %--------------------------
465: A digital curve admits various equivalent definitions \cite{BrimkovKlette}.
466: One of them is the following.
467: A {\em simple closed digital curve} is a set $\rho = \{c_1,c_2,\dots,c_l \}$ of pixels that satisfy
468: the following two axioms:
469: {\rm (A1)} $c_i$ is $\alpha$-adjacent to $c_j$ iff $i = j \pm 1
470: ({\rm modulo} \ l)$, and
471: {\rm (B1)} $\rho$ is one-dimensional with respect to $\alpha$-adjacency, for $\alpha = 0$
472: (Fig. \ref{connect0}a, top-left) or 1 (Fig. \ref{connect0}a, bottom-left).
473: %-----------------
474: To get acquinted with the classical definition of dimension of a digital object the reader
475: is referred to \cite{MYL1971}.
476: For further developments and various results see \cite{KLE2004,BrimkovKlette} and the
477: bibliography therein.
478: For instance, we have the following fact.
479: %-----------
480: \begin{fact}
481: \label{F1}
482: Let a digital object $M$ be one-dimensional with respect to adjacency $\alpha \in \{0,1\}$.
483: Then $M$ does not contain a 2-block.
484: \end{fact}
485: %---------
486:
487: Any connected subset of a closed digital curve is a {\em simple digital arc} (Fig. \ref{connect0}a, right).
488: More in general, by analogy to the classical definition of a curve
489: in the plane\footnote{A curve in $\R^2$ is one-dimensional continuum, where {\em continuum} is any nonempty subset of a ceratin
490: topological space that is compact and topologically connected \cite{Urysohn,Menger}.},
491: a digital curve can be defined as a digital object that is connected and one-dimensional with respect to
492: an adjacency relation $\alpha$ (see Fig. \ref{connect0}b).
493: Applied to digital curves, Theorem \ref{Th1} and Fact \ref{F1} imply the following corollaries.
494:
495: %---------------
496: \begin{corollary}
497: \begin{itemize}
498: \item
499: Let $M$ be a digital curve. Then $t=v-2(p+1-h)$.
500: \item
501: If $M$ is tunnel-free, then $v=2(p+1-h)$.
502: \item
503: If $M$ is a simple digital arc, then $t=v-2(p+1)$.
504: \item
505: If $M$ is a simple tunnel-free digital arc, then $v=2(p+1)$.
506: \item
507: If $M$ is a simple closed digital curve, then $t=v-2p$.
508: \item
509: If $M$ is a simple closed tunnel-free digital curve, then $v=2p$.
510: \end{itemize}
511: \end{corollary}
512: %---------------
513:
514:
515: \section{Final remark}
516: \label{sec:5}
517:
518: In this paper we proposed a formula for the number of tunnels in a digital object.
519: Work in progress is pursuing extension of this result to higher dimensions.
520:
521: %--------------------------
522: \begin{thebibliography}{30}
523:
524: \bibitem{BRI2002}
525: Brimkov, V.E., E.\ Andres, and R.P.\ Barneva,
526: Object discretizations in higher dimensions,
527: {\it Pattern Recognition Letters} {\bf 23} (2002) 623--636
528:
529: \bibitem{BBN}
530: Brimkov, V.E., R.P. Barneva, and Ph. Nehlig, Minimally thin discrete triangulations,
531: In: {\em Volume Graphics}, A. Kaufman, R. Yagel, M. Chen (Eds.),
532: Chapter 3, Springer Verlag, 2000, pp. 51-70
533:
534: \bibitem{BrimkovKlette}
535: Brimkov, V.E. and R. Klette, Curves, hypersurfaces, and good pairs of adjacency relations,
536: {\em Lecture Notes in Computer Science} No 3322, Springer Verlag ( 2004) 270--284
537:
538: \bibitem{CohenKaufman}
539: Cohen-Or, D. and A. Kaufman, 3D line voxelization and connectivity control,
540: {\em IEEE Computer Graphics and Applications} {\bf 17} (6) (1997)
541: 80--87
542:
543: \bibitem{Kaufman}
544: Kaufman, A., D. Cohen, and R. Yagel, Volume graphics,
545: {\em IEEE Computer} {\bf 26}(7) (1993) 51--64
546:
547: \bibitem{KLE2004}
548: Klette, R. and A. Rosenfeld.
549: {\em Digital Geometry - Geometric Methods for Digital Picture
550: Analysis},
551: Morgan Kaufmann, San Francisco, 2004
552:
553: \bibitem{KON2001}
554: Kong, T.Y.,
555: Digital topology,
556: In: Davis, L.S., editor.
557: {\em Foundations of Image Understanding},
558: Kluwer, Boston, Massachusetts, 2001, pp. 33--71
559:
560: \bibitem{Menger}
561: Menger, K., {\em Kurventheorie}, Teubner, Leipzig, Germany, 1932
562:
563: \bibitem{MYL1971}
564: Mylopoulos, J.P. and T.\ Pavlidis,
565: On the topological properties of quantized spaces.
566: I. The notion of dimension,
567: {\em J. ACM} {\bf 18} (1971.) 239--246
568:
569: \bibitem{Rosen}
570: Rosenfeld, A., Arcs and curves in digital pictures,
571: {\em Journal of the ACM} {\bf 18} (1973) 81--87
572:
573: \bibitem{ROS1974a}
574: Rosenfeld, A.,
575: Adjacency in digital pictures, {\em Information and Control}
576: {\bf 26} (1974) 24--33
577:
578: \bibitem{Urysohn}
579: Urysohn, P., \"Uber die allgemeinen Cantorischen Kurven, Annual meeting,
580: Deutsche Mathematiker Vereinigung, Marbourg, Germany, 1923
581:
582: \end{thebibliography}
583:
584: \end{document}
585:
586: