Saved in:
Bibliographic Details
Main Authors: Nguyen, Khoa Tuan, Park, Ho-min, Oh, Gaeun, Vankerschaver, Joris, De Neve, Wesley
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2504.21340
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866909597657202688
author Nguyen, Khoa Tuan
Park, Ho-min
Oh, Gaeun
Vankerschaver, Joris
De Neve, Wesley
author_facet Nguyen, Khoa Tuan
Park, Ho-min
Oh, Gaeun
Vankerschaver, Joris
De Neve, Wesley
contents We propose a novel approach to cervical cell image classification for cervical cancer screening using the EVA-02 transformer model. We developed a four-step pipeline: fine-tuning EVA-02, feature extraction, selecting important features through multiple machine learning models, and training a new artificial neural network with optional loss weighting for improved generalization. With this design, our best model achieved an F1-score of 0.85227, outperforming the baseline EVA-02 model (0.84878). We also utilized Kernel SHAP analysis and identified key features correlating with cell morphology and staining characteristics, providing interpretable insights into the decision-making process of the fine-tuned model. Our code is available at https://github.com/Khoa-NT/isbi2025_ps3c.
format Preprint
id arxiv_https___arxiv_org_abs_2504_21340
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Towards Improved Cervical Cancer Screening: Vision Transformer-Based Classification and Interpretability
Nguyen, Khoa Tuan
Park, Ho-min
Oh, Gaeun
Vankerschaver, Joris
De Neve, Wesley
Computer Vision and Pattern Recognition
Machine Learning
We propose a novel approach to cervical cell image classification for cervical cancer screening using the EVA-02 transformer model. We developed a four-step pipeline: fine-tuning EVA-02, feature extraction, selecting important features through multiple machine learning models, and training a new artificial neural network with optional loss weighting for improved generalization. With this design, our best model achieved an F1-score of 0.85227, outperforming the baseline EVA-02 model (0.84878). We also utilized Kernel SHAP analysis and identified key features correlating with cell morphology and staining characteristics, providing interpretable insights into the decision-making process of the fine-tuned model. Our code is available at https://github.com/Khoa-NT/isbi2025_ps3c.
title Towards Improved Cervical Cancer Screening: Vision Transformer-Based Classification and Interpretability
topic Computer Vision and Pattern Recognition
Machine Learning
url https://arxiv.org/abs/2504.21340