DIMENSIONALITY REDUCTION IN SMALLPOX HISTOPATHOLOGICAL IMAGES USING AUTOENCODER AND KERNEL PCA

Year 2025, Volume: 9 Issue: 2, 331 - 343, 30.08.2025

Nilgün Şengöz , Emine Vargün

https://doi.org/10.46519/ij3dptdi.1708402

Abstract

Histopathological images of smallpox-infected tissue are complex and high-dimensional, which poses challenges for analysis and diagnosis. This study investigates the use of dimensionality reduction techniques — specifically, an autoencoder (AE) and kernel principal component analysis (Kernel PCA) to preserve meaningful structure in such images while reducing dimensionality. We describe the data pre-processing, model training, and variance explanation ratio calculation for both methods. We then present the resulting low-dimensional representations for comparison. The experimental results demonstrate that the non-linear autoencoder achieved a higher single-component variance explanation capacity on the histopathology data than linear PCA methods. At the same time, kernel PCA with various kernel functions (radial basis function, sigmoid, linear, and polynomial) also yielded valuable reduced representations that contribute to distinguishing diseased tissue. Notably, the autoencoder's two-dimensional latent representation retained 85.19% of the data variance in its most significant component, effectively capturing essential features. Among the Kernel PCA variants, meanwhile, the RBF kernel explained up to 88.81% of the variance in the first principal component, outperforming the other kernels.
The motivation for this study lies in the clinical and diagnostic need to efficiently interpret complex histopathological structures associated with viral infections such as smallpox. Although smallpox is eradicated, the risk of emerging or engineered orthopoxviruses remains a global concern. Hence, developing computational tools that can extract discriminative features from such images is not only scientifically relevant but also medically significant for early identification, preparedness, and differential diagnosis of similar conditions. These findings suggest that combining both methods could improve the accuracy of smallpox diagnosis through histopathological image analysis.

Keywords

Dimensionality Reduction , Autoencoder , Kernel PCA , Histopathology , Smallpox , Variance Ratio.

References

• 1. Jolliffe, I.T., Cadima, J., “Principal component analysis: A review and recent developments” Philosophical Transactions of the Royal Society A, Vol. 374, Issue 2065, Pages 20150202, 2016.
• 2. Schölkopf, B., Smola, A., Müller, K.R., “Nonlinear component analysis as a kernel eigenvalue problem”, Neural Computation, Vol. 10, Issue 5, Pages 1299–1319, 1998.
• 3. Bengio, Y., Courville, A., Vincent, P. “Representation learning: a review and new perspectives”, IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol. 35, Issue 8, Pages 1798–1828, 2013. 4. Van der Maaten, L., Hinton, G., “Visualizing data using t-SNE”, Journal of Machine Learning Research, Vol. 9, Pages 2579-2605, 2008.
• 5. McInnes, L., Healy, J., Melville, J., “UMAP: Uniform manifold approximation and projection for dimension reduction”, arXiv preprint arXiv:1802.03426, 2018.
• 6. Shawe-Taylor, J., Cristianini, N., “Kernel methods for pattern analysis”, Cambridge University Press, 2004.
• 7. Nilgün, Ş., “A hybrid approach for detection and classification of sheep-goat pox disease using deep neural networks”, El-Cezerî Journal of Science and Engineering. Vol. 9, Issue 4, Pages 1542-1554, 2022.
• 8. Liu, Y., Chen, P. H., Krause, J., Peng, L., “How to read articles that use machine learning: users’ guides to the medical literature” JAMA, Vol. 322, Issue 18, Pages 1806–1816, 2020.
• 9. Goodfellow, I., Bengio, Y., Courville, A., “Deep learning”, MIT Press. (See Chapter 14 for connections between linear autoencoders and PCA).
• 10. Litjens, G., Kooi, T., Bejnordi, B.E., Setio, A.A. A., Ciompi, F., Ghafoorian, M., Van Der Laak, J.A., Van Ginneken, B., Sánchez, C. I., “A survey on deep learning in medical image analysis”, Medical Image Analysis, Vol. 42, Pages 60-88, 2017.
• 11. Ronneberger, O., Fischer, P., Brox, T., “U-Net: convolutional networks for biomedical image segmentation”, MICCAI, 2015.
• 12. Hinton, G.E., Salakhutdinov, R.R., “Reducing the dimensionality of data with neural networks”, Science, Vol. 313, Issue 5786, Pages 504-507, 2006.
• 13. Kingma, D.P., Welling, M., “Auto-encoding variational bayes”, ICLR, 2014.
• 14. Esteva, A., Kuprel, B., Novoa, R. A., Ko, J., Swetter, S. M., Blau, H. M., Thrun, S., “Dermatologist-level classification of skin cancer with deep neural networks”, Nature, Vol. 542, Issue 7639, Pages 115-118, 2017.
• 15. Cruz-Roa, A., Basavanhally, A., González, F., Gilmore, H., Feldman, M., Ganesan, S., Shih, N., Tomaszewski, J., Madabhushi, A., “Automatic detection of invasive ductal carcinoma in whole slide images with convolutional neural networks”, Medical Imaging 2014”, Digital Pathology, Pages 9041, 2014.