Vision Transformer Embedded Feature Fusion Model with Pre-Trained Transformers for Keratoconus Disease Classification

Keratoconus is a progressive eye disorder that, if undetected, can lead to severe visual impairment or blindness, necessitating early and accurate diagnosis. The primary objective of this research is to develop a feature fusion hybrid deep learning framework that integrates pretrained Convolutional...

Full description

Saved in:
Bibliographic Details
Main Authors: Md Fatin Ishrak, Md Maruf Rahman, Md Imran Kabir Joy, Anna Tamuly, Salma Akter, Dewan M. Tanim, Shahajada Jawar, Nayeem Ahmed, Md Sadekur Rahman
Format: Article
Language:English
Published: Ital Publication 2025-04-01
Series:Emerging Science Journal
Subjects:
Online Access:https://ijournalse.org/index.php/ESJ/article/view/2922
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850279517711499264
author Md Fatin Ishrak
Md Maruf Rahman
Md Imran Kabir Joy
Anna Tamuly
Salma Akter
Dewan M. Tanim
Shahajada Jawar
Nayeem Ahmed
Md Sadekur Rahman
author_facet Md Fatin Ishrak
Md Maruf Rahman
Md Imran Kabir Joy
Anna Tamuly
Salma Akter
Dewan M. Tanim
Shahajada Jawar
Nayeem Ahmed
Md Sadekur Rahman
author_sort Md Fatin Ishrak
collection DOAJ
description Keratoconus is a progressive eye disorder that, if undetected, can lead to severe visual impairment or blindness, necessitating early and accurate diagnosis. The primary objective of this research is to develop a feature fusion hybrid deep learning framework that integrates pretrained Convolutional Neural Networks (CNNs) with Vision Transformers (ViTs) for the automated classification of keratoconus into three distinct categories: Keratoconus, Normal, and Suspect. The dataset employed in this study is sourced from a widely recognized and publicly available online repository. Prior to model development, comprehensive preprocessing techniques were applied, including the removal of low-quality samples, image resizing, rescaling, and data augmentation. The dataset was subsequently partitioned into training, testing, and validation subsets to facilitate robust model training and performance evaluation. Eight state-of-the-art CNN architectures, including DenseNet121, EfficientNetB0, InceptionResNetV2, InceptionV3, MobileNetV2, ResNet50, VGG16, and VGG19, were utilized for feature extraction, while the ViT served as the classification head, leveraging its global attention mechanism for enhanced contextual learning, achieving near-perfect accuracy (e.g., DenseNet121+ViT: 99.28%). This study underscores the potential of hybrid CNN-ViT architectures to revolutionize keratoconus diagnosis, offering scalable, accurate, and efficient solutions to overcome limitations of traditional diagnostic methods while paving the way for broader applications in medical imaging.   Doi: 10.28991/ESJ-2025-09-02-027 Full Text: PDF
format Article
id doaj-art-3e20021ba62244fb966e87f32df536df
institution OA Journals
issn 2610-9182
language English
publishDate 2025-04-01
publisher Ital Publication
record_format Article
series Emerging Science Journal
spelling doaj-art-3e20021ba62244fb966e87f32df536df2025-08-20T01:49:04ZengItal PublicationEmerging Science Journal2610-91822025-04-01921037107510.28991/ESJ-2025-09-02-027824Vision Transformer Embedded Feature Fusion Model with Pre-Trained Transformers for Keratoconus Disease ClassificationMd Fatin Ishrak0Md Maruf Rahman1Md Imran Kabir Joy2Anna Tamuly3Salma Akter4Dewan M. Tanim5Shahajada Jawar6Nayeem Ahmed7Md Sadekur Rahman8Department of Electrical and Computer Engineering, University of Memphis, Memphis,Department of Marketing & Business Analytics, Texas A&M University- Commerce, Texas,MSA in Engineering Management, Central Michigan University, Michigan,Department of Computer Science, University of Memphis, Memphis,Department of Public Administration, Gannon University, Pennsylvania,Department of Computer and Information Science, Gannon University, Pennsylvania,Department of Computer and Information Science, Gannon University, Pennsylvania,Department of Computer Science, University of Memphis, Memphis,Department of Computer Science and Engineering, Daffodil International University, Birulia,Keratoconus is a progressive eye disorder that, if undetected, can lead to severe visual impairment or blindness, necessitating early and accurate diagnosis. The primary objective of this research is to develop a feature fusion hybrid deep learning framework that integrates pretrained Convolutional Neural Networks (CNNs) with Vision Transformers (ViTs) for the automated classification of keratoconus into three distinct categories: Keratoconus, Normal, and Suspect. The dataset employed in this study is sourced from a widely recognized and publicly available online repository. Prior to model development, comprehensive preprocessing techniques were applied, including the removal of low-quality samples, image resizing, rescaling, and data augmentation. The dataset was subsequently partitioned into training, testing, and validation subsets to facilitate robust model training and performance evaluation. Eight state-of-the-art CNN architectures, including DenseNet121, EfficientNetB0, InceptionResNetV2, InceptionV3, MobileNetV2, ResNet50, VGG16, and VGG19, were utilized for feature extraction, while the ViT served as the classification head, leveraging its global attention mechanism for enhanced contextual learning, achieving near-perfect accuracy (e.g., DenseNet121+ViT: 99.28%). This study underscores the potential of hybrid CNN-ViT architectures to revolutionize keratoconus diagnosis, offering scalable, accurate, and efficient solutions to overcome limitations of traditional diagnostic methods while paving the way for broader applications in medical imaging.   Doi: 10.28991/ESJ-2025-09-02-027 Full Text: PDFhttps://ijournalse.org/index.php/ESJ/article/view/2922feature fusion modelkeratoconusvision transformerdensenet121efficientnetb0inceptionresnetv2inceptionv3mobilenetv2resnet50vgg16vgg19.
spellingShingle Md Fatin Ishrak
Md Maruf Rahman
Md Imran Kabir Joy
Anna Tamuly
Salma Akter
Dewan M. Tanim
Shahajada Jawar
Nayeem Ahmed
Md Sadekur Rahman
Vision Transformer Embedded Feature Fusion Model with Pre-Trained Transformers for Keratoconus Disease Classification
Emerging Science Journal
feature fusion model
keratoconus
vision transformer
densenet121
efficientnetb0
inceptionresnetv2
inceptionv3
mobilenetv2
resnet50
vgg16
vgg19.
title Vision Transformer Embedded Feature Fusion Model with Pre-Trained Transformers for Keratoconus Disease Classification
title_full Vision Transformer Embedded Feature Fusion Model with Pre-Trained Transformers for Keratoconus Disease Classification
title_fullStr Vision Transformer Embedded Feature Fusion Model with Pre-Trained Transformers for Keratoconus Disease Classification
title_full_unstemmed Vision Transformer Embedded Feature Fusion Model with Pre-Trained Transformers for Keratoconus Disease Classification
title_short Vision Transformer Embedded Feature Fusion Model with Pre-Trained Transformers for Keratoconus Disease Classification
title_sort vision transformer embedded feature fusion model with pre trained transformers for keratoconus disease classification
topic feature fusion model
keratoconus
vision transformer
densenet121
efficientnetb0
inceptionresnetv2
inceptionv3
mobilenetv2
resnet50
vgg16
vgg19.
url https://ijournalse.org/index.php/ESJ/article/view/2922
work_keys_str_mv AT mdfatinishrak visiontransformerembeddedfeaturefusionmodelwithpretrainedtransformersforkeratoconusdiseaseclassification
AT mdmarufrahman visiontransformerembeddedfeaturefusionmodelwithpretrainedtransformersforkeratoconusdiseaseclassification
AT mdimrankabirjoy visiontransformerembeddedfeaturefusionmodelwithpretrainedtransformersforkeratoconusdiseaseclassification
AT annatamuly visiontransformerembeddedfeaturefusionmodelwithpretrainedtransformersforkeratoconusdiseaseclassification
AT salmaakter visiontransformerembeddedfeaturefusionmodelwithpretrainedtransformersforkeratoconusdiseaseclassification
AT dewanmtanim visiontransformerembeddedfeaturefusionmodelwithpretrainedtransformersforkeratoconusdiseaseclassification
AT shahajadajawar visiontransformerembeddedfeaturefusionmodelwithpretrainedtransformersforkeratoconusdiseaseclassification
AT nayeemahmed visiontransformerembeddedfeaturefusionmodelwithpretrainedtransformersforkeratoconusdiseaseclassification
AT mdsadekurrahman visiontransformerembeddedfeaturefusionmodelwithpretrainedtransformersforkeratoconusdiseaseclassification