Vision Transformer-Enhanced Multi-Descriptor Approach for Robust Age-Invariant Face Recognition

Justice Kwame Appati; Emmanuel Tsifokor; Daniel Kwame Amissah; David Ebo Adjepon-Yamoah

doi:10.1002/ail2.70000

Vision Transformer-Enhanced Multi-Descriptor Approach for Robust Age-Invariant Face Recognition

Justice Kwame Appati, Emmanuel Tsifokor, Daniel Kwame Amissah, David Ebo Adjepon-Yamoah

Department of Computer Science

Research output: Contribution to journal › Article › peer-review

Abstract

This study presents a robust age-invariant face recognition framework, addressing challenges posed by age-related facial variations. Evaluated on the FGNet and Morph II datasets, the system integrates Viola-Jones for face detection, SIFT and LBP for feature extraction, and Vision Transformers (ViTs) for global feature representation. Feature fusion and dimensionality reduction (KPCA, IPCA, UMAP) enhance efficiency while retaining key discriminative information. Using Random Forest, KNN, and XGBoost classifiers, the model achieves 96% accuracy, demonstrating the effectiveness of combining traditional and deep learning techniques in advancing age-invariant face recognition.

Original language	English
Article number	e70000
Journal	Applied AI Letters
Volume	6
Issue number	3
DOIs	https://doi.org/10.1002/ail2.70000
Publication status	Published - Oct 2025

Keywords

age-invariant face recognition
dimensionality reduction
feature extraction
local binary patterns
machine learning
scale-invariant feature transform
vision transformers

Access to Document

10.1002/ail2.70000

Cite this

@article{fdfc1a99a05a4a8ab03c518087e2a975,

title = "Vision Transformer-Enhanced Multi-Descriptor Approach for Robust Age-Invariant Face Recognition",

abstract = "This study presents a robust age-invariant face recognition framework, addressing challenges posed by age-related facial variations. Evaluated on the FGNet and Morph II datasets, the system integrates Viola-Jones for face detection, SIFT and LBP for feature extraction, and Vision Transformers (ViTs) for global feature representation. Feature fusion and dimensionality reduction (KPCA, IPCA, UMAP) enhance efficiency while retaining key discriminative information. Using Random Forest, KNN, and XGBoost classifiers, the model achieves 96\% accuracy, demonstrating the effectiveness of combining traditional and deep learning techniques in advancing age-invariant face recognition.",

keywords = "age-invariant face recognition, dimensionality reduction, feature extraction, local binary patterns, machine learning, scale-invariant feature transform, vision transformers",

author = "Appati, \{Justice Kwame\} and Emmanuel Tsifokor and Amissah, \{Daniel Kwame\} and Adjepon-Yamoah, \{David Ebo\}",

note = "Publisher Copyright: {\textcopyright} 2025 The Author(s). Applied AI Letters published by John Wiley \& Sons Ltd.",

year = "2025",

month = oct,

doi = "10.1002/ail2.70000",

language = "English",

volume = "6",

journal = "Applied AI Letters",

issn = "2689-5595",

publisher = "John Wiley and Sons Inc.",

number = "3",

}

TY - JOUR

T1 - Vision Transformer-Enhanced Multi-Descriptor Approach for Robust Age-Invariant Face Recognition

AU - Appati, Justice Kwame

AU - Tsifokor, Emmanuel

AU - Amissah, Daniel Kwame

AU - Adjepon-Yamoah, David Ebo

PY - 2025/10

Y1 - 2025/10

N2 - This study presents a robust age-invariant face recognition framework, addressing challenges posed by age-related facial variations. Evaluated on the FGNet and Morph II datasets, the system integrates Viola-Jones for face detection, SIFT and LBP for feature extraction, and Vision Transformers (ViTs) for global feature representation. Feature fusion and dimensionality reduction (KPCA, IPCA, UMAP) enhance efficiency while retaining key discriminative information. Using Random Forest, KNN, and XGBoost classifiers, the model achieves 96% accuracy, demonstrating the effectiveness of combining traditional and deep learning techniques in advancing age-invariant face recognition.

AB - This study presents a robust age-invariant face recognition framework, addressing challenges posed by age-related facial variations. Evaluated on the FGNet and Morph II datasets, the system integrates Viola-Jones for face detection, SIFT and LBP for feature extraction, and Vision Transformers (ViTs) for global feature representation. Feature fusion and dimensionality reduction (KPCA, IPCA, UMAP) enhance efficiency while retaining key discriminative information. Using Random Forest, KNN, and XGBoost classifiers, the model achieves 96% accuracy, demonstrating the effectiveness of combining traditional and deep learning techniques in advancing age-invariant face recognition.

KW - age-invariant face recognition

KW - dimensionality reduction

KW - feature extraction

KW - local binary patterns

KW - machine learning

KW - scale-invariant feature transform

KW - vision transformers

UR - https://www.scopus.com/pages/publications/105009987383

U2 - 10.1002/ail2.70000

DO - 10.1002/ail2.70000

M3 - Article

AN - SCOPUS:105009987383

SN - 2689-5595

VL - 6

JO - Applied AI Letters

JF - Applied AI Letters

IS - 3

M1 - e70000

ER -

Vision Transformer-Enhanced Multi-Descriptor Approach for Robust Age-Invariant Face Recognition

Abstract

Keywords

Access to Document

Other files and links

Fingerprint

Cite this