Some properties of overlap and grouping functions and their application to image thresholding

doi:10.1016/j.fss.2012.12.009

Fuzzy Sets and Systems

Volume 229, 16 October 2013, Pages 69-90

https://doi.org/10.1016/j.fss.2012.12.009 Get rights and content

Abstract

In this paper we study under which conditions overlap and grouping functions satisfy some commonly demanded properties such as migrativity or homogeneity. We also recall that the convex combination of overlap (grouping) functions is a new overlap (grouping) function. This property allows us to achieve a consensus between different methods that solve certain problems by means of these functions. We also show one application of this property in image processing.

References (27)

H. Bustince et al.
Ignorance functions. An application to the calculation of the threshold in prostate ultrasound images
Fuzzy Sets Syst.
(2010)
H. Bustince et al.
Overlap functions
Nonlinear Anal.Theory Methods Appl.
(2010)
H. Bustince et al.
Restricted equivalence functions
Fuzzy Sets Syst.
(2006)
H. Bustince et al.
Image thresholding using restricted equivalence functions and maximizing the measures of similarity
Fuzzy Sets Syst.
(2007)
H. Bustince et al.
Migrativity of aggregation operators
Fuzzy Sets Syst.
(2009)
B. Caldairou et al.
A non-local fuzzy segmentation methodapplication to brain MRI
Pattern Recognition
(2011)
E. Castiñeira et al.
Measuring incompatibility between Atanassov's intuitionistic fuzzy sets
Inf. Sci.
(2010)
F. Chiclana et al.
Integrating multiplicative preference relations in a multipurpose decision-making model based on fuzzy preference relations
Fuzzy Sets Syst.
(2001)
F. Durante et al.
A note on the convex combinations of triangular norms
Fuzzy Sets Syst.
(2008)
J. Fodor et al.
On continuous triangular norms that are migrative
Fuzzy Sets Syst.
(2007)

L.K. Huang et al.

Image thresholding by minimizing the measure of fuzziness

Pattern Recognition

(1995)

A. Pradera et al.

A general class of triangular norm-based aggregation operators: quasi-linear T–S operators

Int. J. Approx. Reason.

(2002)

J.A. Sanz et al.

Improving the performance of fuzzy rule-based classification systems with interval-valued fuzzy sets and genetic amplitude tuning

Inf. Sci.

(2010)

Cited by (166)

Two novel three-way decision models based on fuzzy β-covering rough sets and prospect theory under q-rung orthopair fuzzy environments
2024, Expert Systems with Applications
Three-way decision (TWD) and prospect theory (PT) are efficacious theories for investigating uncertain problems. TWD can mitigate the risks associated with rapid decision-making in traditional decision-making. PT can describe the psychological behavior of decision-makers. Furthermore, q-rung orthopair fuzzy sets (q-ROFSs) exhibit wider-ranging applications since they generalize intuitive fuzzy sets and Pythagorean fuzzy sets and unify them under one framework. However, the acquisition of q-ROFSs is an important issue in practical problems. In light of these considerations, this paper develops two TWD models under q-rung orthopair fuzzy environments, employing the principles of PT. To begin, we introduce two novel fuzzy $β$ -covering rough sets that meet the inclusion property. Subsequently, utilizing the proposed fuzzy rough set, we provide an objective methodology to acquire q-ROFSs. Following this, the relative utility function is given through a combination of the value function and EDAS method. Additionally, the weighting function is constructed using two distinct approaches: one that continues to incorporate the ideas of EDAS method, and the other that utilizes the probabilistic dominance $α$ -neighborhood classes to calculate conditional probabilities. After that, classification rules and sorting rules are derived. Ultimately, we substantiate the validity and excellence of the proposed TWD models through comparative and experimental analysis.
Multigranulation fuzzy probabilistic rough sets induced by overlap functions and their applications
2024, Fuzzy Sets and Systems
As we all know, t-norms can be used to construct fuzzy probabilistic rough set (FPRS). Meanwhile, overlap functions (OFs), as a sort of novel aggregation functions different from t-norms, have shown a flourishing situation in terms of applications and theory, especially for the study involving combination of OFs with rough sets. In this paper, we propose a novel OFs-based FPRS named as OFPRS. Specifically, first, we provide a pair of approximation operators of OFPRS via the conditional probability based on OFs. Meanwhile, we present a new OFs-based multigranulation fuzzy probabilistic rough set named as OMGFPRS. Then, we study elementary properties of OFPRS and OMGFPRS. Furthermore, we list practical examples to illustrate the feasibility as well as effectiveness of OFPRS and OMGFPRS, and give a short comparison of the proposed models with existing corresponding FPRS models. Lastly, we develop numerical experiments where OFPRS and OMGFPRS have better classification performance than t-norms-based FPRS.
Construction methods of fuzzy implications on bounded posets
2024, International Journal of Approximate Reasoning
The fuzzy implication on bounded lattices was introduced by Palmeira et al., and the method of extending fuzzy implications on bounded lattices by using retraction was provided. However, we find that the extension of fuzzy implications on bounded lattices can also be realized through homomorphism. In order to get better results, we will continue to study this topic in this paper. In particular, we will focus on the construction methods of fuzzy implications on bounded posets. More precisely, we will give some construction methods of fuzzy implications via $0, 1$ -homomorphism on bounded posets. Then we further study two special kinds of fuzzy implications, $(Q, N)$ -implications and $R_{Q}$ -implications on bounded posets, where $Q$ is a quasi-overlap function. Finally, we discuss the distributive laws and the importation laws of $(Q, N)$ -implications and $R_{Q}$ -implications over a quasi-overlap function $Q$ .
Note on the pseudo-homogeneous overlap and grouping functions
2023, Fuzzy Sets and Systems
In this article, we first prove that the pseudo-homogeneity and quasi-homogeneity are equivalent for overlap and grouping functions, and then give representations of a given pseudo-homogeneous overlap and grouping functions in terms of the single-argument functions induced by the given overlap and grouping functions, respectively. We finally introduce the concept of a pseudo-homogeneous generator pair of overlap function which is applied to construct an overlap function.
Irreducible quasi-D-overlap functions: Matrix representation and diagonal generation
2023, Fuzzy Sets and Systems
As an emerging mathematical model of information aggregation successfully applied in image processing, pattern recognition, decision-making and etc, theoretical research of overlap functions has also continued to develop, especially the research on their related derivative concepts that have emerged in recent years. This work considers so-called irreducible quasi-D-overlap functions defined on grid of unit square, which are one class of discrete aggregation operators originated from overlap functions. The core contents are: (i) it presents the concept of irreducible quasi-D-overlap functions and shows basic properties of them; (ii) it provides equivalent characterization of irreducible quasi-D-overlap functions by means of strong permutation matrixes; (iii) it gives a construction manner for irreducible quasi-D-overlap functions by means of so-called diagonal operators and the greatest and smallest irreducible quasi-D-overlap functions with given diagonal operator are obtained.
Generalizing max pooling via (a,b)-grouping functions for Convolutional Neural Networks
2023, Information Fusion
Due to their high adaptability to varied settings and effective optimization algorithm, Convolutional Neural Networks (CNNs) have set the state-of-the-art on image processing jobs for the previous decade. CNNs work in a sequential fashion, alternating between extracting significant features from an input image and aggregating these features locally through “pooling” functions, in order to produce a more compact representation.
Functions like the arithmetic mean or, more typically, the maximum are commonly used to perform this downsampling operation. Despite the fact that many studies have been devoted to the development of alternative pooling algorithms, in practice, “max-pooling” still equals or exceeds most of these possibilities, and has become the standard for CNN construction.
In this paper we focus on the properties that make the maximum such an efficient solution in the context of CNN feature downsampling and propose its replacement by grouping functions, a family of functions that share those desirable properties. In order to adapt these functions to the context of CNNs, we present $(a, b)$ -grouping functions, an extension of grouping functions to work with real valued data. We present different construction methods for $(a, b)$ -grouping functions, and demonstrate their empirical applicability for replacing max-pooling by using them to replace the pooling function of many well-known CNN architectures, finding promising results.

View all citing articles on Scopus

View full text

Some properties of overlap and grouping functions and their application to image thresholding

Abstract

Fuzzy Sets Syst.

Nonlinear Anal.Theory Methods Appl.

Fuzzy Sets Syst.

Fuzzy Sets Syst.

Fuzzy Sets Syst.

Pattern Recognition

Inf. Sci.

Fuzzy Sets Syst.

Fuzzy Sets Syst.

Fuzzy Sets Syst.

Pattern Recognition

Int. J. Approx. Reason.

Inf. Sci.