Transfer Learning Fusion Approaches for Colorectal Cancer Histopathological Image Analysis

It is well-known that accurate classification of histopathological images is essential for effective diagnosis of colorectal cancer. Our study presents three attention-based decision fusion models that combine pre-trained CNNs (Inception V3, Xception, and MobileNet) with a spatial attention mechanis...

Full description

Saved in:
Bibliographic Details
Main Authors: Houda Saif ALGhafri, Chia S. Lim
Format: Article
Language:English
Published: MDPI AG 2025-06-01
Series:Journal of Imaging
Subjects:
Online Access:https://www.mdpi.com/2313-433X/11/7/210
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849246390593519616
author Houda Saif ALGhafri
Chia S. Lim
author_facet Houda Saif ALGhafri
Chia S. Lim
author_sort Houda Saif ALGhafri
collection DOAJ
description It is well-known that accurate classification of histopathological images is essential for effective diagnosis of colorectal cancer. Our study presents three attention-based decision fusion models that combine pre-trained CNNs (Inception V3, Xception, and MobileNet) with a spatial attention mechanism to enhance feature extraction and focus on critical image regions. A key innovation is the attention-driven fusion strategy at the decision level, where model predictions are weighted by relevance and confidence to improve classification performance. The proposed models were tested on diverse datasets, including 17,531 colorectal cancer histopathological images collected from the Royal Hospital in the Sultanate of Oman and a publicly accessible repository, to assess their generalizability. The performance results achieved high accuracy (98–100%), strong MCC and Kappa scores, and low misclassification rates, highlighting the robustness of the proposed models. These models outperformed individual transfer learning approaches (<i>p</i> = 0.009), with performance differences attributed to the characteristics of the datasets. Gradient-weighted class activation highlighted key predictive regions, enhancing interpretability. Our findings suggest that the proposed models demonstrate the potential for accurately classifying CRC images, highlighting their value for research and future exploration in diagnostic support.
format Article
id doaj-art-3929e8c0099b49bc99ef40cd263bb5de
institution Kabale University
issn 2313-433X
language English
publishDate 2025-06-01
publisher MDPI AG
record_format Article
series Journal of Imaging
spelling doaj-art-3929e8c0099b49bc99ef40cd263bb5de2025-08-20T03:58:31ZengMDPI AGJournal of Imaging2313-433X2025-06-0111721010.3390/jimaging11070210Transfer Learning Fusion Approaches for Colorectal Cancer Histopathological Image AnalysisHouda Saif ALGhafri0Chia S. Lim1Department of Information Technology, College of Computing and Information Sciences, University of Technology and Applied Sciences, Muscat 133, OmanGraduate School of Technology, Asia Pacific University of Technology and Innovation, Kuala Lumpur 57000, MalaysiaIt is well-known that accurate classification of histopathological images is essential for effective diagnosis of colorectal cancer. Our study presents three attention-based decision fusion models that combine pre-trained CNNs (Inception V3, Xception, and MobileNet) with a spatial attention mechanism to enhance feature extraction and focus on critical image regions. A key innovation is the attention-driven fusion strategy at the decision level, where model predictions are weighted by relevance and confidence to improve classification performance. The proposed models were tested on diverse datasets, including 17,531 colorectal cancer histopathological images collected from the Royal Hospital in the Sultanate of Oman and a publicly accessible repository, to assess their generalizability. The performance results achieved high accuracy (98–100%), strong MCC and Kappa scores, and low misclassification rates, highlighting the robustness of the proposed models. These models outperformed individual transfer learning approaches (<i>p</i> = 0.009), with performance differences attributed to the characteristics of the datasets. Gradient-weighted class activation highlighted key predictive regions, enhancing interpretability. Our findings suggest that the proposed models demonstrate the potential for accurately classifying CRC images, highlighting their value for research and future exploration in diagnostic support.https://www.mdpi.com/2313-433X/11/7/210colorectal cancerdecision fusiontransfer learningspatial attention mechanismshistopathological images
spellingShingle Houda Saif ALGhafri
Chia S. Lim
Transfer Learning Fusion Approaches for Colorectal Cancer Histopathological Image Analysis
Journal of Imaging
colorectal cancer
decision fusion
transfer learning
spatial attention mechanisms
histopathological images
title Transfer Learning Fusion Approaches for Colorectal Cancer Histopathological Image Analysis
title_full Transfer Learning Fusion Approaches for Colorectal Cancer Histopathological Image Analysis
title_fullStr Transfer Learning Fusion Approaches for Colorectal Cancer Histopathological Image Analysis
title_full_unstemmed Transfer Learning Fusion Approaches for Colorectal Cancer Histopathological Image Analysis
title_short Transfer Learning Fusion Approaches for Colorectal Cancer Histopathological Image Analysis
title_sort transfer learning fusion approaches for colorectal cancer histopathological image analysis
topic colorectal cancer
decision fusion
transfer learning
spatial attention mechanisms
histopathological images
url https://www.mdpi.com/2313-433X/11/7/210
work_keys_str_mv AT houdasaifalghafri transferlearningfusionapproachesforcolorectalcancerhistopathologicalimageanalysis
AT chiaslim transferlearningfusionapproachesforcolorectalcancerhistopathologicalimageanalysis