1. Introduction
Let
(
resp. ) represent the set of integers (
resp. natural numbers), and
be the
n times Cartesian product of
,
. Besides, let
(
resp. ) be the set of even (
resp. odd) natural numbers. Digital geometry mainly deals with discrete objects in
from the viewpoints of digital
k-curve and digital
k-surface theory, where the
k-adjacency means the digital
k-connectivity of
(see (2.1) in
Section 2).
Let
be the category consisting of the set of digital images
, denoted by
, and the set of
k-continuous maps of
, denoted by
(for more details see
Section 2). Motivated by the study of homotopy fixed point sets in Reference [
1], we are recently interested in the set of cardinal numbers of fixed point sets of all
k-continuous maps of a digital image
[
2,
3] because this topic can be used in studying some “motions of a rigid body with a fixed point” [
4,
5] from the viewpoint of digital geometry. To be specific, given a digital image
in
, a paper [
2] explored some features of (
k-homotopy) fixed point sets related to this issue in a
setting. Then, the authors of Reference [
2] used the terminology, “fixed point spectrum” denoted by
. However, a recent paper [
3] changed the term “spectrum” into “alignment” because the term “spectrum” has been used in the field of functional analysis as a very popular and important notion related to the bounded or unbounded liner operator [
6]. Indeed, the “alignment” of a fixed point sets in Reference [
3] is considered to be a digital image with 2-adjacency (see Definition 4) instead of only a set [
2], which appears a little bit difference between them.
For a digital image
, let
and
. Besides, we will denote the cardinality of a set
X with
. In addition, we use the notation “
” for introducing a new terminology. Then, with the following setting (for more details see
Section 3)
we will take the notation
for brevity. In particular, this paper explores some conditions supporting the perfectness of
because the notion of perfectness can play an important role in mechatronics and digital geometry.
A recent paper [
7] proved that a
k-isomorphism preserves a
k-homotopy, a
k-homotopy equivalence,
k-contractibility (for more details see Theorem 2 and Corollaries 1 and 2 of Reference [
7]). This finding can facilitate fixed point theory and homotopy theory in a
setting. Given a simple closed
k-curve with
l elements in
, denoted by
, we proved that
is perfect if and only if
is
k-contractible [
3]. Although Reference [
2] formulated
for the case of
, we need to improve it. Indeed, we can further consider it for the case of
(for more details see Lemma 2 in this paper). Concretely, given
, we can formulate
without any limitation of
l of
. With this approach, the following queries can be raised.
- (Q1)
Given with , under what condition does the set have both the elements and ?
Indeed, this issue plays an important role in studying .
Let us now denote the
m-iterated digital wedge of
with
, (for more details see
Section 2 in the present paper). Up to now, in digital geometry there is no research of the fixed point sets of
, where
. After formulating
for the case of
, given a number
l which is either even or odd, we may raised he following queries (see Remark 1).
- (Q2)
How can we establish ?
- (Q3)
How many 2-components are there in ?
- (Q4)
Are there some relationships among the numbers m, l, and the perfectness of ?
- (Q5)
Given a simple k-path with a length d, what conditions make perfect?
- (Q6)
How can we formulate ?
- (Q7)
Is a digital k-homotopy invariant?
- (Q8)
Are there certain relationships between the perfectness of and that of ?
The rest of the paper is organized as follows:
Section 2 recalls basic backgrouds and some properties related to the study of fixed points of digital images as well as the multiplicative property of a digital fundamental group [
8,
9].
Section 3 explores a condition that makes
perfect. In particular, given
with
, we will propose a certain condition of which
contains both the elements
and
. Indeed, this finding strongly facilitates examining if
is perfect.
Section 4 proves that
is perfect if and only if
. Besides, we prove that
has two 2-components if and only if
.
Section 5 begins with
, where
. Then, it investigates some properties of
for the case
,
. Besides, after joining a simple
k-path
onto
W to produce a digital wedge with a
k-adjacency, denoted by
, we prove that
is perfect if and only if
, where
d is the length of
P. Finally, we investigate a certain condition that makes
perfect,
. In particular,
is proved not to be a digital
k-homotopy invariant.
Section 6 expands the obtained results associated with
of
in
Section 4 and
Section 5 into the cases of which
l is odd and
.
Section 7 concludes the paper and refers to a further work.
2. Preliminaries
A pair
consisting of a set
and a certain
k-adjacency of
was initially called a
digital image,
[
10,
11]. After that, Reference [
12] firstly generalized this approach into the high dimensional digital image
with one of the
k-adjacency relations of
. To study
in a
setting,
, the following digital
k-adjacency (or digital
k-connectivity) was taken in Reference [
12] (see also Reference [
13]), as follows:
For a natural number
t,
, the distinct points
are
-adjacent if at most
t of their coordinates differ by
and the others coincide.
According to this approach, the
-adjacency relations of
, are formulated [
12] (see also Reference [
13]) as follows:
For instance, we obtain the following [
12,
13]:
Using these
k-adjacency relations of
of (1),
, we will call
a digital image on
,
. Besides, these
k-adjacency relations can be essential for studying digital products with normal adjacencies [
9,
12] and calculating digital
k-fundamental groups of digital products [
9,
12] (see Theorem 2 in this paper). For
with
, the set
with 2-adjacency is called a digital interval [
14,
15].
Using a digital
k-adjacency of
, we observe that a digital image
is a digital space [
16] (see also Reference [
7]). Hereafter,
is assumed in
, with one of the
k-adjacency of (1). Let us recall some terminology, notions, and backgrouds need for this study [
8,
10,
11,
12,
14,
15].
Assume
with
. Then, by a
k-path with
elements in
X we means the sequence
such that
and
are
k-adjacent if
[
15].
We say that
is
k-connected if for any distinct points
there is a
k-path
in
X such that
and
(for more details see Reference [
7]). Besides, a singleton set is assumed to be
k-connected (for more details see Reference [
7]).
Given a digital image
, by the
k-component of
, we mean the maximal
k-connected subset of
containing the point
x [
15].
By a simple
k-path from
x to
y in
, we mean the finite set
such that
and
are
k-adjacent if and only if
, where
and
[
15]. Then, the length of this set
is denoted by
.
A simple closed
k-curve (or simple
k-cycle) with
l elements in
, denoted by
[
12,
15],
, is defined to be the set
such that
and
are
k-adjacent if and only if
. Then, the number
l of
depends on both the dimension
n of
and the
k-adjacency (for details, see Remark 1). Hereafter we use the notation
to abbreviate
.
Remark 1. Let us investigate how the number l of is taken. According to the k-adjacency of , that is, of (1), we observe that the number l of is determined, as follows:
(Case 1) In the case , according to the notion of , it is clear that the number l should be an even number . For instance, let us consider , and , .
(Case 2) In view of the concept of , it is obvious that no exists. For instance, neither nor exists. However, the existence of depends on both k and n. To be precise, the number of depends on the situation. For instance, while we can take , no exists.
(Case 3) In the case , as to , if , then it is clear that the number l can be even or odd (see Figure 1a). For instance, consider , , and (see Figure 1a (1)–(3)). In general, and are considered depending on the dimension , and . In addition, is admissible. For instance, consider , , and so on. Owing to Remark 1, in terms of the number
l of
, the following properties of are observed.
For instance, we can consider
,
,
,
and so on. Hereafter, regarding
l of
, we will follow the property (2).
Given
and a point
, the following notion of “digital
k-neighborhood of
with radius 1” is defined, as follows [
12]:
For more general cases of
, see References [
7,
12]. The digital
-continuity of a map
in Reference [
11] can be represented by using the
k-neighborhood in (3), as follows:
Proposition 1. [12] A function is (digitally) -continuous if and only if for every , . In Proposition 1, in the case
, the map
f is called a ‘
k-continuous’ map to abbreviate the
-continuity of the given map
f. Using the digital continuity of maps between two digital images, let us recall the category
DTC consisting of the following two pieces of data [
12], called the digital topological category, as follows:
The set of , where , as objects of DTC denoted by ;
For every ordered pair of objects , the set of all -continuous maps between them as morphisms of DTC, denoted by .
In DTC, for the case , we will use the notation DTC(k).
In some literature, since there is a certain confusion of characterizing both a k-path and a -continuous map, we need to confirm some difference between them, as follows:
Remark 2. Given a k-path on , we may establish a -continuous map whose image of by f is considered to be the given k-path, that is, as a sequence with . However, given a -continuous map , not every image by the given map f as a sequence, that is, the sequence , is always a k-path on (see the notion of k-path in the previous part) because some of the points and can be equal. Of course, depending on the situation, based on this sequence , we can take a k-path as a subsequence of .
To compare digital images up to similarity, we often use the notion of a -isomorphism (or k-isomorphism), as follows:
Definition 1. [8] (-homeomorphism in Reference [17]) Consider two digital images and in and , respectively. Then, a map is called a -isomorphism if h is a -continuous bijection and further, is -continuous. Then, we use the notation . In the case , the map h is called a k-isomorphism. Let us now recall the notions of a digital wedge which can be used in studying fixed point sets from the viewpoint of digital geometry. Given two digital images
and
, a
digital wedge, denoted by
, is initially defined [
8,
12] as the union of the digital images
and
(for more details see
Figure 1,
Figure 2,
Figure 3 and
Figure 4), where
- (1)
is a singleton, say .
- (2)
and
are not
k-adjacent, where two sets
and
are said to be
k-adjacent if
and there are at least two points
and
such that
a is
k-adjacent to
b [
14].
- (3)
is k-isomorphic to and is k-isomorphic to (see Definition 1).
When studying digital wedges in a
setting, we are strongly required to follow this approach. Besides,
might be considered to be
. Indeed, this digital wedge is quite different from the classical one point union (or wedge) in typical topology [
18] and standard graph theory [
19] by the
k-adjacency referred to in (2) above.
A digital topological version of the strong graph adjacency of a product of two typical graphs [
20] was initially developed in Reference [
12]. It is called a normal adjacency of digital product [
8]. Indeed, this notion is strongly related to the calculation of digital fundamental groups of digital products [
9] and further, an automorphism group of a digital covering space of a digital product [
9] (see Theorem 2 and Corollary 1 below).
Motivated by the pointed digital
k-homotopy in References [
21,
22] (see also Reference [
17]), the concept of
k-homotopy relative to a subset
is established, as follows:
Definition 2. [12,17] Let and be a digital image pair and a digital image in and , respectively. Let be -continuous functions. Suppose there exist and a function such that - (1)
for all and ;
- (2)
for all , the induced function given by for all is -continuous.
- (3)
for all , the induced function given by
for all is -continuous; Then, H is said to be a -homotopy between f and g [17]. - (4)
Furthermore, for all , for all and for all [12].
Then, we call H a -homotopy relative to A between f and g, and f and g are said to be -homotopic relative to A in Y, in symbols [12]. In Definition 2, if
, then we say that
F is a pointed
-homotopy at
[
17]. In the case
and
, we call a
k-homotopy to abbreviate
-homotopy. If, for some
,
is
k-homotopic to the constant map in the space
X relative to
, then
is said to be
pointed k-contractible [
17].
Based on this
k-homotopy, the notion of digital homotopy equivalence firstly introduced in Reference [
23], as follows:
Definition 3. [23] Given two digital images and , if there are k-continuous maps and such that the composite is k-homotopic to and the composite is k-homotopic to , then the map is called a k-homotopy equivalence and is denoted by . Besides, is said to be k-homotopy equivalent to . If the identity map is k-homotopy equivalent to a certain constant map , we say that is k-contractible. In particular, in the case and , the map h is called a -homotopy equivalence. With the several concepts such as digital
k-homotopy class [
21,
22], Khalimsky operation of two
k-homotopy classes [
21], trivial extension [
17], Reference [
17] defined the digital
k-fundamental group, denoted by
. Also, we have the following: If
X is pointed
k-contractible, then it is clear that
is a trivial group [
17]. Using the homotopy lifting theorem, and the unique digital lifting theorem [
12], we obtain the following [
12]:
Theorem 1. [12] (1) For a non-k-contractible , is an infinite cyclic group. (2) For two non-k-contractible spaces , is a free group with two generators of which they have infinite orders.
For instance, is a free group with two generators, .
Given two digital images
in
, consider the digital product
in
. For every point
, if there is a
k-adjacency of
(see (1)) such that
, then we say that the
k-adjacency for
is
C-compatible [
9]. Since both the
k-fundamental group and the
k-contractibility of a digital image
are so related to the study of
(see Theorem 4), we need to recall the following properties.
Theorem 2. [9] Assume that is a simple closed -curve with elements in , and they are not -contractible. If there is a C-compatible k-adjacency of its digital product , then the k-fundamental group is isomorphic to . Theorem 2 strongly facilitates the classification of digital covering spaces [
12], as follows:
Corollary 1. [9] Assume a digital product with a C-compatible k-adjacency. Let be a radius 2--covering map, , where and . Then, the automorphism group is isomorphic to the group . 3. Existence of a Perfect Alignment of Fixed Point Sets
This section explores some conditions that make an alignment of fixed point sets of a digital image 2-connected (or perfect). As usual, we say that a
digital topological property is a property of a digital image
which is invariant under digital
k-isomorphisms. As mentioned in the previous part, this paper takes the notation
to highlight the set of
k-continuous self-maps of
, as follows:
where
Using the set in (4), we define the following:
Definition 4. [3] Given , is said to be an alignment of fixed point sets of . A paper [
2] only used the notation
as just a set without the 2-adjacency. In Definition 4, we remind that the pair
is assumed to be a digital image with 2-adjacency as a subset of
.
Definition 5. [3] Given , if , then (or for brevity) is said to be perfect. This notion can play an important role in studying motions of rigid objects with fixed points in the field of robotics.
Theorem 3. In , is a digital topological property [2,3]. Proof. For two
k-isomorphic images
and
, since we have
(see Corollary 4.4 of Reference [
2]), the proof is completed. ☐
Given , we obviously obtain the following:
Lemma 1. (1) [2,3]. (2) [3]. (3) For , [2]. (4) In the case , .
References [
2,
3] (or arXiv version of [
2]) studied
only for the case of even number
l (see Lemma 1). After recalling Remark 1, let us now formulate
for an odd number
l of
.
Lemma 2. For an odd number l of , , that is, .
Proof. Regarding the assertion, it is clear that
because only a singleton digital image has the fixed point property [
11,
24] (for more details see Remark 10 of [
7]), a constant map is also a
k-continuous map, and the identity map of
is obviously a
k-continuous self-map of
. Thus, it suffices to consider certain
k-continuous self-maps
f of
such that
, that is,
. Then, this set
should be a certain
k-path which is not a simple
k-cycle (or a certain simple closed
k-curve) on
. More precisely, consider the map (see
Figure 2)
defined by
Then, this map
f is a
k-continuous self-map of
such that
(see (6) with Proposition 1), where
. In view of Proposition 1, the image by the map
proposed in (6) has the maximal number of the set (for instance, see
Figure 2a–c)
Indeed, for each number
, it is clear that there are many types of
k-continuous self-maps
g of
depending on the number
l such that
By (5), (7), and (8), we conclude that . ☐
Example 1. (1) . To be precise, for convenience, let us consider the self-map f of in Figure 2a with . Then, regarding , according to the map in (6), consider the map Then, it is clear that f is an 8-continuous map with . Similarly, for any , we can have another several 8-continuous self-maps g of such that . Hence we have because it is clear that .
Similarly, we obtain the following:
- (2)
(see Figure 2b). - (3)
(see Figure 2c).
For a digital image
which is
k-connected and
, it is clear that
[
2] and further, by Lemmas 1 and 2, and the property (2), we obtain the following: For any
l of
and
of
,
In (9)(2), the set need not be a subset of . For instance, and . Besides, using Lemma 1(3) and Lemma 2, we observe .
Given
with
, we need to check if there is the number
. Indeed, a paper [
2] studied this property with the following lemma (see Lemma 4.8 of Reference [
2]).
Lemma 3. [2] Let X be connected with . Then, if and only if there are distinct points with . Unlike the notations
and
in Lemma 3, in digital topology we have ordinarily used the following notations: For a given digital image
, using the digital
k-connectivity of (1), we are used to take the following notations [
11,
12,
14]. According to the literature, for
, we recall the notations [
11,
12,
14]
- (1)
;
- (2)
;
- (3)
(see the notation in (3)); and
- (4)
We may represent of (3) above in such a way that
, that is, the notation of of (3).
The notations
and
in Lemma 3 are defined, as follows:
A recent paper [
3] contains some incorrect part in Remark 3 of Reference [
3]. Indeed, the remark was based on the preprint posted in arXiv “Fixed point sets in Digital topology, 1(2019), 1-25” which is the preprint version of Reference [
2]. Then, there was not any introduction of
and
in this preprint. With the notion in (10), the current version of Lemma 3 above is correct anyway (disregard the lines 4–11 from the bottom of the page 11 of Reference [
3]). But, owing to the unusual notations of
and
in Lemma 3, a reader can invoke some confusion. Thus, after following the standard notations referred to in (1)–(4) above, we can avoid some potential confusion. Thus, we had better follow the popular and standard notations as stated in (1)–(4) above to finally obtain the following representation of Lemma 3.
Lemma 4 (Representation of Lemma 3). Let be k-connected with . Then, if and only if there are distinct points with .
Remark 3. Unlike Lemma 1(3), using Lemma 2, we can deal with , . For insance, we obtain .
Since the notion of
k-retract also play an important role in studying
, let us now recall it, as follows: For
, we say that
is a
k-retract of
[
17] if there is a
k-continuous map
such that for any
,
. Then, we call the map
r a
k-retraction from
onto
.
As mentioned in
Section 2, the notion of
k-contractibility of
strongly contributes to the study of
. Indeed, the following 6-contractibility of the given digital image in Lemma 5 has been often used in digital topology [
7,
24]. To be precise, given a digital image
with
, in order to examine if
belongs to
, Reference [
2] has used the following property.
Lemma 5. [7,24] Let be the digital unit cube with 6-adjacency. Then, it is 6-contractible. To be specific, though
is 6-contractible, it is clear that
[
2]. However, we need to point out that the 6-contractibility of this 3-dimensional digital cube
was already proven in References [
7,
24] (see Theorem 2.6(1) of Reference [
24] and Remark 2(2) of Reference [
7]).
Since each of
,
,
, and
, plays an important role in digital topology (see Theorem 1), let us intensively explore alignments of the fixed point sets of them, for example,
As a representation of Lemma 1 of Reference [
3], for any
l of
we obtain the following (see Remark 1):
Theorem 4. is perfect if and only if is trivial.
Proof. Using “Contrapositive law”, let us suppose that is not trivial. Then, by Theorem 1, it is obvious that is not k-contractible. Namely, by Remark 1, we obtain at least . In this case, by Lemmas 1(3) and 2, we need to consider the following two cases:
(Case 1) In the case l is even, by Lemma 1(3), we obtain . Hence, if , then is not perfect because (see the difference between and l in Lemma 1(3)).
(Case 2) In the case l is odd, by Lemma 2, we have . Hence, if , then is not perfect because (see the difference between and l).
Conversely, if is trivial, by Theorem 1, we obtain . For instance, we can see the cases , , and . Thus, by Theorem 1 and Lemma 1, we clearly obtain which is perfect. ☐
By Theorem 4, it is clear that is perfect if and only if . Using Lemmas 3 and 4, we obtain the following result which can play an important role in investigating the perfectness of .
Lemma 6. Given which is k-connected, assume that there are two distinct points such that . Then, we obtain the following:
- (1)
is k-connected.
- (2)
is a k-retract of .
Proof. In the case
, the proof is straightforward because a singleton
is
k-connected (for more details see the notion of
k-disconnectedness of Reference [
7]). Thus, we may assume
.
(1) With the hypothesis “
”, using “Reductio ad Absurdum”, let us suppose that
is not
k-connected. Then, there are non-empty sets
(for details see the notion of
k-disconnectedness in Reference [
7]) such that
Then, there are at least two points such that because is k-connected. Owing to the hypothesis “”, each and should be in . Then, owing to the condition (11), we have a contradiction.
For instance, see the cases in
Figure 1b(1)–(2). To be specific, see the two points
and
in
Figure 1b(1) and the two points
and
in
Figure 1b(2) which support this assertion.
(2) With the hypothesis, let us consider the map
defined by
Then, by Proposition 1, with the hypothesis of this theorem we obtain
Thus, r is a k-continuous map such that , where is the restriction of the map r to the given set A. Hence r is a k-retraction from onto . ☐
Lemma 7. [2] Let be a k-retract of . Then, . When investigating the perfectness of a given digital image, we need the following assertion which answers to the question .
Theorem 5. Let be k-connected with . If there are three or four distinct points such that and further,then, . Proof . (Case 1) With the hypothesis of (1) of (13), we prove the assertion. By Lemmas 3 and 4, owing to the hypothesis
, it is clear
. Furthermore, in
, owing to the hypothesis that there are certain two points
in
such that
by Lemmas 4 and 6(1), we obtain
because
. Owing to the hypothesis
, by Lemma 6(2), we see that
is a
k-retract of
. Furthermore, by Lemma 7, we also obtain
so that both
and
belong to
.
(Case 2) With the hypothesis of (2) of (13), we prove the assertion. By Lemmas 3 and 4, owing to the hypothesis
, it is clear that
. Furthermore, in
, owing to the hypothesis that there are two distinct points
in
such that
, by Lemmas 4 and 6, we obtain
because
. Owing to the condition that
, by Lemma 6(2), we see that
is a
k-retract of
. Then, by Lemma 7, we conclude that both
and
belong to
. For instance, see the cases in
Figure 1b(1) and (2) and
Figure 3b(1) and (2). ☐
Example 2. (1) For any , we now confirm that (see (8) of Reference [3])has at least the elements because there are at least three points in that satisfies the condition (1) of (13) and For instance, consider the case in Figure 3a. Then, we obtain that has at least the elements because there are at least three points in that satisfies the condition (1) of (13) and (2) By Lemmas 2, 3, and 4, since does not have the number 15 (see Figure 3f), we conclude that is not perfect. (3) By Theorem 5, we obtain .
Remark 4. The two conditions (1) and (2) of Theorem 5 are independent.
Proof. Let us consider the two digital images
and
with
in
Figure 1c. Then, by Lemma 1(3) and Theorem 5, we see that
Naively, each and are perfect so that the numbers belong to .
For
, since there are the three points
,
and
in
X such that
Hence satisfies the condition (1) of Theorem 5. However, it is clear that does not satisfy the condition (2) of Theorem 5.
Meanwhile, let us now consider the case
in
Figure 1c(2). Then, there are the following four points
in
Y such that
Besides, it is clear that this does not satisfy the condition (1) of Theorem 5. ☐
Remark 5. The converse of Theorem 5 does not hold.
Proof. Let
in
Figure 1c(1). Naively, consider
in
Figure 1c(1). Then, by Lemmas 1(3) and 4,
and we also have
so that it is clear that
because
. However, the digital image
does not satisfy both of the conditions (1) and (2) of Theorem 5. For instance, in
, while there are points
such that
, there are no points satisfying either the condition (1) or the condition (2) of (13) in Theorem 5. Namely,
does not satisfies each of the conditions (1) and (2) of (13). ☐
Owing to Theorem 5, the following is obtained:
Corollary 2. Given with , if there is a set with such that Then, .
Example 3. (1) Given a simple k-path with , we obtain .
(2) Let . Then, as stated in Example 2(3), which is perfect.
(3) Let be a digital wedge of some simple k-paths. Then, is perfect.
4. Formulation of
This section initially formulates . Then, the following queries are naturally raised (see (Q2) and (Q3)).
(4-1) Under what conditions is perfect ?
(4-2) How many 2-components are there in ?
Since the exploration of the perfectness or non-perfectness of
and plays an important role in digital geometry, this section intensively studies this topic and its generalization. In particular, we prove that is perfect if and only if . Besides, we prove that has two 2-components if and only if , as follows:
Theorem 6. For , .
Proof. Though there are many kinds of k-continuous self-maps f of , regarding , it suffices to consider only the maps f such that
- (a)
; or
- (b)
; or
- (c)
; or
- (d)
f does not have any fixed point of it, where means the restriction function f to the given set A.
Firstly, from (a), since
has the cardinality
, using Lemma 1(3), we have
Secondly, from (b), using Lemma 1(3), we obtain
Thirdly, from (c)–(d), we have
After comparing the three numbers
,
, and
from (15)–(17), we conclude that
☐
By Theorem 6, the following is obtained.
Theorem 7. (1) For , is perfect if and only if .
(2) For , has two 2-components if and only if .
Proof. (1) Let us now take the difference from (18)
From (19), if , then the perfectness of holds. Thus, we obtain the following: is perfect if and only if .
(2) In (19) above, count on the case , that is, . In view of (18), we conclude that has the two 2-components and in the digital image if and only if . ☐
By Theorems 6 and 7, we obtain the following:
Example 4. (1) which is not perfect.
(2) has two 2-components which is not perfect.
(3) has two 2-components, which is not perfect.
(4) has two 2-components, which is not perfect.
Using the property of (18), with the property (2), we obtain the following:
Theorem 8. If , then .
Before proving this theorem, we need to point out that this formulation of in this assertion holds for . More precisely, for the case is already stated in Lemma 1(3). As a generalized case of , for the only case , we now prove the assertion. Hence we will prove this assertion by using “transfinite induction” for .
Proof . (Step 1) In the case
, for convenience, let
and
and
, that is,
. Regarding
, it is sufficient to consider the following
k-continuous self-maps
f of
.
According to (20), we now investigate with the following three cases.
Firstly, according to (20)(1) and (2), by Lemma 1(3), we obtain
Secondly, according to (20)(3), we have
For instance, consider a 26-continuous self-map
f of
in
Figure 4 such that
Based on this case, we obtain seven types of 26-continuous self-maps of such that .
Thirdly, according to (20)(4), we obtain
Hence, owing to these three subsets from (21), (22), and (23) as subsets of
, we obtain
(Step 2) In the case
(see Theorem 6), using a method similar to the consideration of (20) and further, following a certain procedure similar to (21)–(23), we obtain
Thus, using Lemma 1(3) (see also (Step 1) above) and the properties of (24) and (25), we obtain the following.
By transfinite induction for
, from (26) we obtain
☐
As mentioned in the earlier part, the formula of (27) also holds for (see Lemma 1(3)).
Corollary 3. If , then the cardinality of is equal to .
5. Perfectness of , Where is a Simple -Path and
This section initially formulates both and
, where is a simple k-path and . To proceed with this work, first of all we need to investigate a certain condition that makes perfect, . In the case , we already mentioned that is perfect if and only if (see Lemma 1 and Theorem 4). Motivated by Theorem 8, we obtain the following:
Theorem 9. In the case , is perfect if and only if .
Proof. From (27), take the difference between the two numbers
and
, that is,
Indeed, the quantity in (28) plays a crucial role in investigating cardinalities of the fixed point sets of in . According to the numbers m and l, based on the difference in (28), if , that is, , then we obtain the following:
is perfect if and only if . ☐
By Theorem 9, for the case
it turns out that
is not 2-connected. Thus, we now address the query (Q5) stated in
Section 1. Before dealing with (Q5), we need to remind the following feature which can play a crucial role in addressing the query (Q5) and study
from the viewpoint of digital geometry.
Remark 6. According to (28), given , regardless of the number of m of the given m-iterated digital wedge of , the difference of (28) is constant. Thus, the perfectness of is equivalent to that of .
Motivated by Theorems 6, 7, and 9, after joining a simple k-path onto to produce a digital wedge with the k-adjacency, we investigate a certain condition that makes perfect, as follows:
Theorem 10. In the case , is perfect if and only if , where is a simple k-path and d is the length of P.
Proof. From (28), if , that is, , then
is 2-connected and further, the converse also holds. Thus, we complete the proof. ☐
Using Theorems 7 and 9, let us now investigate a certain condition that makes perfect.
By Lemma 6 and Theorems 5 and 9, since is perfect, we prove the following:
Theorem 11. In the case , is perfect if and only if .
Before proving this assertion, in the case , we recall that the assertion obviously holds (see (18)).
Proof. By Theorem 5, it is clear that
is perfect. Naively, we have (see Example 2(3))
because the cardinality of
is
. With the property (28) of Theorem 9, owing to the inequality
, that is,
, since
, we have the following:
is perfect () if and only if . ☐
Example 5. (1) In the case , is perfect.
(2) In the case , is perfect.
As a general case of Theorem 11, the following is obtained.
Theorem 12. In the case , is perfect if and only if .
Proof. From (28), if , that is, , then we obtain is perfect. Besides, the converse also holds. ☐
Remark 7. The obtained results from Theorems 9, 10, 11, and 12 are independent from the given k-adjacency.
Definition 6. A property P of digital images is called a “digital k-homotopy property” (or k-homotopy invariant) provided that it is preserved by all digital k-homotopy equivalences.
To be precise, a certain property P is a digital k-homotopy property if and only if, for an arbitrary k-homotopy equivalence , that has P implies that also has P.
Proposition 2. is not a digital k-homotopy invariant.
Proof. As a counterexample, consider the two digital images
and
in
Figure 1c(2). Though they are 8-homotopy equivalent, their alignments of the fixed point sets of them are different, as follows:
so that
. ☐
6. For an Odd Number , Characterization of
As mentioned in (2), since no
exists, in this section we are only interested in the number
of
. Unlike the study of
preceded in
Section 4 and
Section 5, this section studies
for the case that
l is odd instead of even, that is,
(see Remark 1 and the property (2)). Then, Lemma 2 plays a crucial role in proceeding with this work. Thus, first of all, let us compare the two assertions in Lemmas 1 and 2.
Remark 8. Given , according to the two numbers or in Lemmas 1 and 2, we obtain the following:
(Case 1) In the case (see Lemma 1), we have (Case 2) In the case (see Lemma 2), we have Thus, we see that these two alignments of the given and have the same cardinality ‘’. However, we observe some difference between them from Case 1 and Case 2. To be precise, let us count on the two differencesfrom the above two alignments of the given fixed point sets of and . Depending on the choice of or , we have the corresponding differences “” or “a” as stated in (29), respectively. This observation also plays an important role in studying or comparing the two cases of according to the choice of or . As mentioned in Remark 1, to characterize , we need to consider the case because no exists.
Theorem 13. If and ,then .
Before proving the assertion, regarding , we need to follow the property (2).
Proof. Using Lemma 2 and a method similar to the proof of Theorem 8, we prove the assertion. To be precise, for the case we have already stated in Lemma 2. As a generalized case of , for the only case , we now prove the assertion. Hence we will prove this assertion by using “transfinite induction” for .
(Step 1) In the case
, let
and
and
, that is,
(see
Figure 3g as an example). Regarding
, motivated by the method of (20), it suffices to consider the following
k-continuous self-maps
f of
.
According to (30), we now investigate with the following three cases.
Firstly, according to (30)(1) and (2), by Lemma 2, we obtain
Secondly, according to (30)(3), we have
because
(see Lemma 2).
For instance, consider an 8-continuous self-map
f of
in
Figure 3g such that for each subset
of
so that we obtain
.
With this approach, we obtain several types of 8-continuous self-maps of such that .
Thirdly, according to (30)(4), we obtain
Hence, owing to these three sets from (31), (32), and (33) as subsets of
, we obtain
(Step 2) In the case
, using a method similar to the consideration of (30) and further, following a certain procedure similar to (31)–(33), we obtain
Thus, using Lemma 2 (see also (Step 1) above) and using several properties similar to the properties of (31), (32), and (33), we we obtain the following.
By transfinite induction for
, from (36) we obtain
☐
Example 6. .
Remark 9. After comparing (27) and (37), we have the different types of alignments of depending on the number l which is either even or odd.
Corollary 4. In the case , is not perfect.
Proof. By Remark 1, Lemma 2, and (37), the proof is completed because . ☐
Example 7. Using the properties (2) and (3) of (36), by Theorem 13, we obtain the following:
- (1)
, which is not perfect.
- (2)
, which is perfect.
- (3)
, which is not perfect.
- (4)
, which is not perfect.
From (6.9), let us take the difference between
and
, that is,
This quantity is essential to studying the perfectness of , where is a certain simple k-path. In view of (37) and (38), we obtain the following:
Corollary 5. For , has two 2-components.
Proof. In view of (38), has two 2-components if and only , that is, . Since no exists, only for , the assertion holds. ☐
Theorem 14. In the case , is perfect if and only if .
Proof. By Theorem 5, it is clear that
is perfect (see Example 2(3)). Naively, we have
because the cardinality of
is
. Withe the property (38), owing to the inequality
, that is,
, we have the following:
is perfect if and only if . ☐
Example 8. (1) In the case , is perfect.
(2) In the case , is perfect.
Remark 10. According to (38), given , regardless of the number of m of the given m-iterated digital wedge of , the difference of (38) is constant. Thus, the perfectness of is equivalent to that of .
As a general case of Theorem 14, the following is obtained.
Theorem 15. In the case , is perfect if and only if .
Proof. From (38), by Remark 1, we obtain that is perfect if and only , that is, . ☐
Theorem 16. In the case , is perfect if and only if , where is a simple k-path and d is the length of P.
Proof. From (38), we obtain that is perfect if and only if , that is, . ☐