Advanced Methodologies in Cervical Cancer Classification using Deep Learning and Ensemble Techniques

Cervical cancer remains a major cause of morbidity among women globally, particularly in regions with limited healthcare infrastructure. Despite advancements in screening techniques, manual cytological interpretation of Pap smear images is subject to observer bias and diagnostic inconsistency. This...

Full description

Saved in:
Bibliographic Details
Published in2025 3rd International Conference on Self Sustainable Artificial Intelligence Systems (ICSSAS) pp. 1225 - 1231
Main Authors Duary, Kanishka, Rajgarhia, Jigyasha, D., Viji
Format Conference Proceeding
LanguageEnglish
Published IEEE 11.06.2025
Subjects
Accuracy
Cervical cancer
Cervical cancer classification
Computer architecture
Deep learning
Fuzzy ensemble
Hybrid CNN
Measurement
Medical services
Microprocessors
Observers
Pap smear image analysis
Robustness
Transfer learning
Online AccessGet full text
DOI10.1109/ICSSAS66150.2025.11081310

Cover

More Information
Summary:Cervical cancer remains a major cause of morbidity among women globally, particularly in regions with limited healthcare infrastructure. Despite advancements in screening techniques, manual cytological interpretation of Pap smear images is subject to observer bias and diagnostic inconsistency. This paper introduces a hybrid deep learning framework integrating InceptionV3, MobileNetV2, and Inception ResNetV2, fused with a fuzzy-distance ensemble method to improve reliability and accuracy in cervical cancer cell classification. Our approach builds upon existing works in deep cytology analysis and addresses limitations in generalization and class imbalance found in prior research. The system demonstrates superior performance on the SiPaKMeD dataset, showing strong potential for real-world clinical deployment.
DOI:10.1109/ICSSAS66150.2025.11081310