Detecting basal cell carcinoma in skin histopathological images using deep learning

In this thesis we explore different machine learning techniques that are common in image classification to detect the presence of Basal Cell Carcinoma (BCC) in digital skin histological images. Since digital histology images are extremely large, we first focused on determining the presence of BCC at the patch level, using pre-trained deep convolutional neural networks as feature extractors to compensate for the size of our datasets. The experimental results show that our patch level classifiers obtained an area under the receiver operating characteristic curve (AUC) of 0.981. Finally, we used our patch classifiers to generate a bag of scores for a given whole slide image (WSI), and attempted multiple ways of combining these scores to produce a single significant score to predict the presence of BCC in the given WSI. Our best performing model obtained an AUC of 0.991 in 86 samples of digital skin biopsies, 43 of which had BCC.

Name thesis:degree_name

Master

Department dc:contributor.department

Massachusetts Institute of Technology. Department of Electrical Engineering and Computer Science

Grantor dc:publisher

Massachusetts Institute of Technology

Year dc:date.issued

2018

Author dc:creator

• Calderón Gómez, Tomás Alberto.

Advisor dc:contributor.advisor

• Tomaso Poggio.

• Electrical Engineering and Computer Science.

Statement dc:rights

• MIT theses are protected by copyright. They may be viewed, downloaded, or printed from this source but further reproduction or distribution in any format is prohibited without written permission.

Licence dc:rights.uri

• http://dspace.mit.edu/handle/1721.1/7582

Language dc:language.iso

eng

Handle dc:identifier.uri

https://hdl.handle.net/1721.1/121624

OAI identifier oai:identifier

oai:dspace.mit.edu:1721.1/121624