We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

eess.IV

Change to browse by:

References & Citations

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Electrical Engineering and Systems Science > Image and Video Processing

Title: Generalized Categorisation of Digital Pathology Whole Image Slides using Unsupervised Learning

Abstract: This project aims to break down large pathology images into small tiles and then cluster those tiles into distinct groups without the knowledge of true labels, our analysis shows how difficult certain aspects of clustering tumorous and non-tumorous cells can be and also shows that comparing the results of different unsupervised approaches is not a trivial task. The project also provides a software package to be used by the digital pathology community, that uses some of the approaches developed to perform unsupervised unsupervised tile classification, which could then be easily manually labelled.
The project uses a mixture of techniques ranging from classical clustering algorithms such as K-Means and Gaussian Mixture Models to more complicated feature extraction techniques such as deep Autoencoders and Multi-loss learning. Throughout the project, we attempt to set a benchmark for evaluation using a few measures such as completeness scores and cluster plots.
Throughout our results we show that Convolutional Autoencoders manages to slightly outperform the rest of the approaches due to its powerful internal representation learning abilities. Moreover, we show that Gaussian Mixture models produce better results than K-Means on average due to its flexibility in capturing different clusters. We also show the huge difference in the difficulties of classifying different types of pathology textures.
Subjects: Image and Video Processing (eess.IV); Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG)
Cite as: arXiv:2012.13955 [eess.IV]
  (or arXiv:2012.13955v1 [eess.IV] for this version)

Submission history

From: Mostafa Ibrahim [view email]
[v1] Sun, 27 Dec 2020 14:38:22 GMT (12873kb,D)

Link back to: arXiv, form interface, contact.