Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Explainable Contrastive and Cost-Sensitive Learning for Cervical Cancer Classification

Ashfiqun Mustari, Rushmia Ahmed, Afsara Tasnim, Jakia Sultana Juthi, G M Shahariar

TL;DR

The paper tackles automated cervical cancer cell classification from Pap smear images and the need for trustworthy, interpretable models. It integrates five pretrained CNNs with cost-sensitive learning and supervised contrastive learning, supplemented by explainable AI techniques to reveal decision reasons. On the SIPaKMeD dataset, the approach achieves up to 97.29% accuracy, surpassing several prior methods. By addressing class imbalance and提供 interpretable visual explanations via GradCAM and LIME, the work demonstrates a pathway toward reliable AI-assisted cervical cancer screening applicable in resource-limited settings.

Abstract

This paper proposes an efficient system for classifying cervical cancer cells using pre-trained convolutional neural networks (CNNs). We first fine-tune five pre-trained CNNs and minimize the overall cost of misclassification by prioritizing accuracy for certain classes that have higher associated costs or importance. To further enhance the performance of the models, supervised contrastive learning is included to make the models more adept at capturing important features and patterns. Extensive experimentation are conducted to evaluate the proposed system on the SIPaKMeD dataset. The experimental results demonstrate the effectiveness of the developed system, achieving an accuracy of 97.29%. To make our system more trustworthy, we have employed several explainable AI techniques to interpret how the models reached a specific decision. The implementation of the system can be found at - https://github.com/isha-67/CervicalCancerStudy.

Explainable Contrastive and Cost-Sensitive Learning for Cervical Cancer Classification

TL;DR

The paper tackles automated cervical cancer cell classification from Pap smear images and the need for trustworthy, interpretable models. It integrates five pretrained CNNs with cost-sensitive learning and supervised contrastive learning, supplemented by explainable AI techniques to reveal decision reasons. On the SIPaKMeD dataset, the approach achieves up to 97.29% accuracy, surpassing several prior methods. By addressing class imbalance and提供 interpretable visual explanations via GradCAM and LIME, the work demonstrates a pathway toward reliable AI-assisted cervical cancer screening applicable in resource-limited settings.

Abstract

This paper proposes an efficient system for classifying cervical cancer cells using pre-trained convolutional neural networks (CNNs). We first fine-tune five pre-trained CNNs and minimize the overall cost of misclassification by prioritizing accuracy for certain classes that have higher associated costs or importance. To further enhance the performance of the models, supervised contrastive learning is included to make the models more adept at capturing important features and patterns. Extensive experimentation are conducted to evaluate the proposed system on the SIPaKMeD dataset. The experimental results demonstrate the effectiveness of the developed system, achieving an accuracy of 97.29%. To make our system more trustworthy, we have employed several explainable AI techniques to interpret how the models reached a specific decision. The implementation of the system can be found at - https://github.com/isha-67/CervicalCancerStudy.
Paper Structure (18 sections, 2 equations, 6 figures, 4 tables)

This paper contains 18 sections, 2 equations, 6 figures, 4 tables.

Table of Contents

  1. Introduction
  2. Related Works
  3. Methodology
  4. Input
  5. Image Preprocessing
  6. Training
  7. Models
  8. Cost-Sensitive Learning
  9. Supervised Contrastive Learning
  10. Performance Evaluation
  11. Interpretation
  12. Evaluation
  13. Hyper-Parameter Setting
  14. Experimental Results
  15. Comparison with Existing Works
  16. ...and 3 more sections

Figures (6)

  • Figure 1: Proposed Methodology
  • Figure 2: Steps of Supervised Contrastive Learning
  • Figure 3: Five sample outputs of correctly classified instances of DenseNet-169 using gradient-based XAI techniques
  • Figure 4: Five sample outputs of misclassified instances of DenseNet-169 using gradient-based XAI techniques
  • Figure 5: Five sample outputs of correctly classified instances of DenseNet-169 using Perturbation-based Visualization technique LIME
  • ...and 1 more figures