Automated classification of the histopathology and frozen section images of prostate cancer using transfer learning framework

Date

2016

Authors

Kaur, Arleen

Journal Title

Journal ISSN

Volume Title

Publisher

Abstract

Frozen section biopsy is commonly used in oncological surgery to perform rapid microscopic analysis of a specimen when immediate answers are required. Currently, pathologists manually grade the prostate tissues biopsies using the Gleason grading schema. The frozen section procedure is challenging and requires the patient to take an appointment with the pathologist at least a day prior to the surgery (www.ncbi.nlm.nih.gov/pmc/articles/PMC3347896/) providing enough time for tissue analysis. Although frozen section procedure provides a rapid diagnosis, there are limitations to this such as poor image quality and limited data availability. The poor image quality makes them inadequate for feature extraction and the application of the existing algorithms are not effective enough for correct classification and grading. The state of art method of classification assumes that the testing and training images should share the same feature space and distribution. However, in some cases, the training and testing data used for classification purposes may be from domains. To solve this, if the knowledge transfer could be successfully done, then it would advance the performance of the system by avoiding exclusive data labeling efforts.

The goals of this thesis are to design a CAD system utilizing both, the frozen and normal prostate cancer biopsy images. Including: (1) classify the normal and frozen prostate biopsy images either as cancerous or non-cancerous; (2) classify the normal and frozen prostate cancer biopsy images into their respective grades; (3) develop an automated system that combines the classification of frozen and normal biopsy sections; and (4) analyze the performance of the presented classification system using different evaluating parameters.

Description

This item is available only to currently enrolled UTSA students, faculty or staff. To download, navigate to Log In in the top right-hand corner of this screen, then select Log in with my UTSA ID.

Keywords

Citation

Department

Electrical and Computer Engineering