Fashion image classification using fuzzy logic

Zineb Ibnou Cheikh; Alexander Gegov; Alexandar Ichtev

doi:10.1109/ICARAI67046.2025.11137853

Fashion image classification using fuzzy logic

Zineb Ibnou Cheikh, Alexander Gegov, Alexandar Ichtev

Technical University of Sofia

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Abstract

This paper explores the use of a Mamdani fuzzy inference system for binary image classification on the Fashion-MNIST dataset, distinguishing T-shirts from other clothing items. Using three interpretable features, average pixel intensity, vertical symmetry, and horizontal symmetry, the fuzzy model is compared against K-Nearest Neighbours (KNN) and Multilayer Perceptron (MLP). While the fuzzy model has lower overall accuracy, it achieves significantly higher recall for the T-shirt class and offers greater interpretability, supporting its value in explainable AI.

Original language	English
Title of host publication	ICARAI 2025 - International Conference Automatics, Robotics and Artificial Intelligence, Proceedings
Publisher	Institute of Electrical and Electronics Engineers Inc.
Number of pages	7
ISBN (Electronic)	9781665465663
ISBN (Print)	9781665465670
DOIs	https://doi.org/10.1109/ICARAI67046.2025.11137853
Publication status	Published - 3 Sept 2025
Event	3rd International Conference Automatics, Robotics and Artificial Intelligence, ICARAI 2025 - Sozopol, Bulgaria Duration: 13 Jun 2025 → 15 Jun 2025

Conference

Conference	3rd International Conference Automatics, Robotics and Artificial Intelligence, ICARAI 2025
Country/Territory	Bulgaria
City	Sozopol
Period	13/06/25 → 15/06/25

Keywords

explainable AI
Fashion-MNIST
fuzzy logic
image classification
Mamdani system

Access to Document

10.1109/ICARAI67046.2025.11137853

Fuzzy_Logic_7_Accepted author manuscript (Post-print), 469 KBLicence: CC BY

Cite this

@inproceedings{0349a02d48ce451aa464615ee89dca16,

title = "Fashion image classification using fuzzy logic",

abstract = "This paper explores the use of a Mamdani fuzzy inference system for binary image classification on the Fashion-MNIST dataset, distinguishing T-shirts from other clothing items. Using three interpretable features, average pixel intensity, vertical symmetry, and horizontal symmetry, the fuzzy model is compared against K-Nearest Neighbours (KNN) and Multilayer Perceptron (MLP). While the fuzzy model has lower overall accuracy, it achieves significantly higher recall for the T-shirt class and offers greater interpretability, supporting its value in explainable AI.",

keywords = "explainable AI, Fashion-MNIST, fuzzy logic, image classification, Mamdani system",

author = "Cheikh, \{Zineb Ibnou\} and Alexander Gegov and Alexandar Ichtev",

note = "No ISSN; 3rd International Conference Automatics, Robotics and Artificial Intelligence, ICARAI 2025 ; Conference date: 13-06-2025 Through 15-06-2025",

year = "2025",

month = sep,

day = "3",

doi = "10.1109/ICARAI67046.2025.11137853",

language = "English",

isbn = "9781665465670",

booktitle = "ICARAI 2025 - International Conference Automatics, Robotics and Artificial Intelligence, Proceedings",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

address = "United States",

}

Cheikh, ZI, Gegov, A & Ichtev, A 2025, Fashion image classification using fuzzy logic. in ICARAI 2025 - International Conference Automatics, Robotics and Artificial Intelligence, Proceedings. Institute of Electrical and Electronics Engineers Inc., 3rd International Conference Automatics, Robotics and Artificial Intelligence, ICARAI 2025, Sozopol, Bulgaria, 13/06/25. https://doi.org/10.1109/ICARAI67046.2025.11137853

TY - GEN

T1 - Fashion image classification using fuzzy logic

AU - Cheikh, Zineb Ibnou

AU - Gegov, Alexander

AU - Ichtev, Alexandar

N1 - No ISSN

PY - 2025/9/3

Y1 - 2025/9/3

N2 - This paper explores the use of a Mamdani fuzzy inference system for binary image classification on the Fashion-MNIST dataset, distinguishing T-shirts from other clothing items. Using three interpretable features, average pixel intensity, vertical symmetry, and horizontal symmetry, the fuzzy model is compared against K-Nearest Neighbours (KNN) and Multilayer Perceptron (MLP). While the fuzzy model has lower overall accuracy, it achieves significantly higher recall for the T-shirt class and offers greater interpretability, supporting its value in explainable AI.

AB - This paper explores the use of a Mamdani fuzzy inference system for binary image classification on the Fashion-MNIST dataset, distinguishing T-shirts from other clothing items. Using three interpretable features, average pixel intensity, vertical symmetry, and horizontal symmetry, the fuzzy model is compared against K-Nearest Neighbours (KNN) and Multilayer Perceptron (MLP). While the fuzzy model has lower overall accuracy, it achieves significantly higher recall for the T-shirt class and offers greater interpretability, supporting its value in explainable AI.

KW - explainable AI

KW - Fashion-MNIST

KW - fuzzy logic

KW - image classification

KW - Mamdani system

UR - https://www.scopus.com/pages/publications/105017544971

U2 - 10.1109/ICARAI67046.2025.11137853

DO - 10.1109/ICARAI67046.2025.11137853

M3 - Conference contribution

AN - SCOPUS:105017544971

SN - 9781665465670

BT - ICARAI 2025 - International Conference Automatics, Robotics and Artificial Intelligence, Proceedings

PB - Institute of Electrical and Electronics Engineers Inc.

T2 - 3rd International Conference Automatics, Robotics and Artificial Intelligence, ICARAI 2025

Y2 - 13 June 2025 through 15 June 2025

ER -

Fashion image classification using fuzzy logic

Abstract

Conference

Keywords

Access to Document

Fingerprint

Cite this