View in EDS

DeepIFSim: A deep-learning and intuitionistic fuzzy similarity framework for face identification.

Saved in:
Bibliographic Details
Title: DeepIFSim: A deep-learning and intuitionistic fuzzy similarity framework for face identification.
Authors: Khan, Muhammad Jabir1 (AUTHOR) jabirkhan.uos@gmail.com, Al-Kenani, Ahmad N.2 (AUTHOR) analkenani@kau.edu.sa, Ding, Weiping1,3 (AUTHOR) dwp9988@163.com
Source: Knowledge-Based Systems. Jul2026, Vol. 346, pN.PAG-N.PAG. 1p.
Subjects: Deep learning, Biometric identification, Machine learning
Abstract: This study presents a novel closed-set face identification framework in which intuitionistic fuzzy similarity measures are integrated with deep feature representations extracted from pre-trained neural networks. First, new sine-based similarity measures for intuitionistic fuzzy sets are introduced, and their mathematical properties are thoroughly analyzed. These measures are then incorporated into a training-free identification framework, termed DeepIFSim, where facial images are represented by deep embeddings and modeled using intuitionistic fuzzy representations to capture uncertainty in feature distributions. Unlike conventional deep-metric-learning approaches, which require extensive network training and large labeled datasets, the proposed framework operates directly on pre-trained deep features and performs identification by maximizing similarity between probe and gallery samples. This design enables efficient deployment in scenarios with limited computational resources or training data. Experimental evaluations conducted on widely used face datasets, namely the Labeled Faces in the Wild (LFW) and Olivetti Research Laboratory (ORL) datasets, demonstrate that the proposed method achieves competitive identification performance compared with classical similarity-based methods, statistical classifiers, and representative deep-learning baselines. In particular, DeepIFSim attains identification accuracies of 99.72% and 98.04% on the ORL and LFW datasets, respectively. The results highlight that the integration of intuitionistic fuzzy modeling with deep feature representations provides a flexible and interpretable alternative to traditional deep-learning pipelines while maintaining strong recognition performance. These characteristics make DeepIFSim a practical solution for closed-set face identification tasks in real-world applications. • Novel face identification framework using intuitionistic fuzzy similarity measures. • Sine-based fuzzy similarity functions with proven mathematical properties. • DeepIFSim integrates IFS measures with features from pre-trained deep learning models. • Validated on face (LFW & ORL), butterfly, and flower datasets with high accuracy. • No training required, reproducible on Colab, adaptable to many domains. [ABSTRACT FROM AUTHOR]
Copyright of Knowledge-Based Systems is the property of Elsevier B.V. and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)
Database: Engineering Source
Description
Abstract:This study presents a novel closed-set face identification framework in which intuitionistic fuzzy similarity measures are integrated with deep feature representations extracted from pre-trained neural networks. First, new sine-based similarity measures for intuitionistic fuzzy sets are introduced, and their mathematical properties are thoroughly analyzed. These measures are then incorporated into a training-free identification framework, termed DeepIFSim, where facial images are represented by deep embeddings and modeled using intuitionistic fuzzy representations to capture uncertainty in feature distributions. Unlike conventional deep-metric-learning approaches, which require extensive network training and large labeled datasets, the proposed framework operates directly on pre-trained deep features and performs identification by maximizing similarity between probe and gallery samples. This design enables efficient deployment in scenarios with limited computational resources or training data. Experimental evaluations conducted on widely used face datasets, namely the Labeled Faces in the Wild (LFW) and Olivetti Research Laboratory (ORL) datasets, demonstrate that the proposed method achieves competitive identification performance compared with classical similarity-based methods, statistical classifiers, and representative deep-learning baselines. In particular, DeepIFSim attains identification accuracies of 99.72% and 98.04% on the ORL and LFW datasets, respectively. The results highlight that the integration of intuitionistic fuzzy modeling with deep feature representations provides a flexible and interpretable alternative to traditional deep-learning pipelines while maintaining strong recognition performance. These characteristics make DeepIFSim a practical solution for closed-set face identification tasks in real-world applications. • Novel face identification framework using intuitionistic fuzzy similarity measures. • Sine-based fuzzy similarity functions with proven mathematical properties. • DeepIFSim integrates IFS measures with features from pre-trained deep learning models. • Validated on face (LFW & ORL), butterfly, and flower datasets with high accuracy. • No training required, reproducible on Colab, adaptable to many domains. [ABSTRACT FROM AUTHOR]
ISSN:09507051
DOI:10.1016/j.knosys.2026.116141