Inter-observer Reliability in Computer-aided Diagnosis of Diabetic Retinopathy

João Gonçalves, Teresa Conceição, Filipe Soares

2019

Abstract

The rapid growth of digital data in healthcare demands medical image analysis to be faster, precise and, at the same time, decentralized. Deep Learning (DL) fits well in this scenario, as there is an enormous data to sift through. Diabetic Retinopathy (DR) is one of the leading causes of blindness that can be avoided if detected in early stages. In this paper, we aim to compare the agreement of different machine learning models against the performance of highly trained ophthalmologists (human graders). Overall results show that transfer learning in the renowned CNNs has a strong agreement even in different datasets. This work also presents an objective comparison between classical feature-based approaches and DL for DR classification, specifically, the interpretability of these approaches. The results show that Inception-V3 CNN was indeed the best-tested model across all the performance metrics in distinct datasets, but with lack of interpretability. In particular, this model reaches the accuracy of 89% on the EyePACS dataset.

Paper Citation

in Harvard Style

Gonçalves J., Conceição T. and Soares F. (2019). Inter-observer Reliability in Computer-aided Diagnosis of Diabetic Retinopathy. In Proceedings of the 12th International Joint Conference on Biomedical Engineering Systems and Technologies (BIOSTEC 2019) - Volume 5: HEALTHINF; ISBN 978-989-758-353-7, SciTePress, pages 481-491. DOI: 10.5220/0007580904810491

in Bibtex Style

@conference{healthinf19,
author={João Gonçalves and Teresa Conceição and Filipe Soares},
title={Inter-observer Reliability in Computer-aided Diagnosis of Diabetic Retinopathy},
booktitle={Proceedings of the 12th International Joint Conference on Biomedical Engineering Systems and Technologies (BIOSTEC 2019) - Volume 5: HEALTHINF},
year={2019},
pages={481-491},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0007580904810491},
isbn={978-989-758-353-7},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 12th International Joint Conference on Biomedical Engineering Systems and Technologies (BIOSTEC 2019) - Volume 5: HEALTHINF
TI - Inter-observer Reliability in Computer-aided Diagnosis of Diabetic Retinopathy
SN - 978-989-758-353-7
AU - Gonçalves J.
AU - Conceição T.
AU - Soares F.
PY - 2019
SP - 481
EP - 491
DO - 10.5220/0007580904810491
PB - SciTePress