[Paper Review] A Comparative Analysis of CNN-Based Pretrained Models for the Detection and Prediction of Monkeypox

Sourav Saha, Trina Chakraborty+2 more|arXiv (Cornell University)|Jan 20, 2023

Poxvirus research and outbreaks10 citations

TL;DR

The paper evaluates multiple CNN-based pretrained models and a majority voting ensemble for monkeypox detection across several class pairs, reporting notable performance for certain models and the ensemble.

ABSTRACT

Monkeypox is a rare disease that raised concern among medical specialists following the convi-19 pandemic. It's concerning since monkeypox is difficult to diagnose early on because of symptoms that are similar to chickenpox and measles. Furthermore, because this is a rare condition, there is a knowledge gap among healthcare professionals. As a result, there is an urgent need for a novel technique to combat and anticipate the disease in the early phases of individual virus infection. Multiple CNN-based pre-trained models, including VGG-16, VGG-19, Restnet50, Inception-V3, Densnet, Xception, MobileNetV2, Alexnet, Lenet, and majority Voting, were employed in classification in this study. For this study, multiple data sets were combined, such as monkeypox vs chickenpox, monkeypox versus measles, monkeypox versus normal, and monkeypox versus all diseases. Majority voting performed 97% in monkeypox vs chickenpox, Xception achieved 79% in monkeypox against measles, MobileNetV2 scored 96% in monkeypox vs normal, and Lenet performed 80% in monkeypox versus all.

Motivation & Objective

• Address the urgent need for automated tools to aid early monkeypox detection and prediction.
• Compare the performance of multiple CNN-based pretrained models on monkeypox-related classification tasks.
• Identify which architectures and ensemble methods yield the strongest discrimination between monkeypox and similar diseases.

Proposed method

• Apply CNN pretrained models (VGG-16, VGG-19, Restnet50, Inception-V3, Densnet, Xception, MobileNetV2, Alexnet, Lenet) to monkeypox classification tasks.
• Use majority voting as an ensemble technique to aggregate model decisions.
• Evaluate models on datasets comprising monkeypox vs chickenpox, monkeypox vs measles, monkeypox vs normal, and monkeypox vs all diseases.
• Report classification performance for each model and scenario.
• Leverage standard CNN-based transfer learning without detailing novel architectural changes.

Experimental results

Research questions

• RQ1 Which CNN-based pretrained architectures best distinguish monkeypox from chickenpox, measles, and normal conditions?
• RQ2 Does an ensemble majority voting approach improve monkeypox detection across different disease pairings?
• RQ3 How do common architectures (e.g., VGG, Inception, Xception, MobileNet) perform relative to each other on monkeypox classification tasks?
• RQ4 What are the relative strengths and limitations of pretrained CNNs for detecting monkeypox in comparisons to similar diseases?

Key findings

• Majority voting achieves 97% accuracy for monkeypox vs chickenpox.
• Xception achieves 79% accuracy for monkeypox vs measles.
• MobileNetV2 achieves 96% accuracy for monkeypox vs normal.
• Lenet achieves 80% accuracy for monkeypox vs all.
• Other models show varying performance across the different class pairings across the datasets.
• The study highlights the potential of pretrained CNNs and ensemble methods for rapid monkeypox detection.