Case-based reasoning of a deep learning network for prediction of early stage of oesophageal cancer

Gao, Xiaohong W. ORCID: https://orcid.org/0000-0002-8103-6624, Braden, Barbara, Zhang, Leishi ORCID: https://orcid.org/0000-0002-3158-2328, Taylor, Stephen, Pang, Wei and Petridis, Miltos (2020) Case-based reasoning of a deep learning network for prediction of early stage of oesophageal cancer. Expert Update: Spring 2020 (Vol. 20, No. 1). In: 24th UK Symposium on Case-Based Reasoning (UKCBR 2019), 17 Dec 2019, Cambridge, UK. . ISSN 1465-4091 [Conference or Workshop Item]

PDF - Final accepted version (with author's formatting) Restricted to Repository staff and depositor only Download (406kB)

Abstract

Case-Based Reasoning (CBR) is a form of analogical reasoning in which the information for a (new) query case is determined based on the known cases in a database with established information. CBR has now become part of artificial intelligence. While deep machine learning techniques have demonstrated state of the art results in many fields, their transparency status of those hidden layers have cast double in many applications, especially in the medical field, where clinicians need to know the reasons of decision making delivered by a computer system. This study aims to provide a visual explanation while performing classification of endoscopic oesophageal videos. Instead of generating a different model to explain the predictions given by a deep learning architecture as having been conducted by many studies, which employ varying priors, this work integrates the interpretation and decision-making together by producing a set of profiles that in appearance resemble the training samples and hence explain the outcome of classification. Furthermore, different from many explainable networks that highlight key regions or points of the input that activate the network, this work is based on whole training images i.e. case-based, where each training image belongs to one of the classes. Preliminary results have demonstrated the classification accuracy of 95 % for training and 75% for testing while applying 500 training data (with 10% for testing split randomly) for each of three classes of `cancer', `high grade' and `suspicious' of oesophageal squamous cancer from endoscopy videos. Future work includes collection of large annotated data set and improving classification accuracy.

Actions (login required)

View Item

Statistics

DownloadsShow export options

Export as CSVXMLJSON

Activity Overview

6 month trend

11Downloads

6 month trend

188Hits

Additional statistics are available via IRStats2.