Automated diagnosis of retinopathy by content-based image retrieval

Edward Chaum, Thomas P. Karnowski, V. Priya Govindasamy, Mohamed Abdelrahman, Kenneth W. Tobin

Research output: Contribution to journalArticle

55 Citations (Scopus)

Abstract

PURPOSE: To describe a novel computer-based image analysis method that is being developed to assist and automate the diagnosis of retinal disease. METHODS: Content-based image retrieval is the process of retrieving related images from large database collections using their pictorial content. The content feature list becomes the index for storage, search, and retrieval of related images from a library based upon specific visual characteristics. Low-level analyses use feature description models and higher-level analyses use perceptual organization and spatial relationships, including clinical metadata, to extract semantic information. RESULTS: We defined, extracted, and tested a large number of region- and lesion-based features from a dataset of 395 retinal images. Using a statistical hold-one-out method, independent queries for each image were submitted to the system and a diagnostic prediction was formulated. The diagnostic sensitivity for all stratified levels of age-related macular degeneration ranged from 75% to 100%. Similarly, the sensitivity of detection and accuracy for proliferative diabetic retinopathy ranged from 75% to 91.7% and for nonproliferative diabetic retinopathy, ranged from 75% to 94.7%. The overall purity of the diagnosis (specificity) for all disease states in the dataset was 91.3%. CONCLUSIONS: The probabilistic nature of content-based image retrieval permits us to make statistically relevant predictions regarding the presence, severity, and manifestations of common retinal diseases from digital images in an automated and deterministic manner.

Original languageEnglish (US)
Pages (from-to)1463-1477
Number of pages15
JournalRetina
Volume28
Issue number10
DOIs
StatePublished - Nov 1 2008

Fingerprint

Retinal Diseases
Diabetic Retinopathy
Macular Degeneration
Semantics
Libraries
Databases
Datasets
Metadata

All Science Journal Classification (ASJC) codes

  • Ophthalmology

Cite this

Chaum, E., Karnowski, T. P., Govindasamy, V. P., Abdelrahman, M., & Tobin, K. W. (2008). Automated diagnosis of retinopathy by content-based image retrieval. Retina, 28(10), 1463-1477. https://doi.org/10.1097/IAE.0b013e31818356dd

Automated diagnosis of retinopathy by content-based image retrieval. / Chaum, Edward; Karnowski, Thomas P.; Govindasamy, V. Priya; Abdelrahman, Mohamed; Tobin, Kenneth W.

In: Retina, Vol. 28, No. 10, 01.11.2008, p. 1463-1477.

Research output: Contribution to journalArticle

Chaum, E, Karnowski, TP, Govindasamy, VP, Abdelrahman, M & Tobin, KW 2008, 'Automated diagnosis of retinopathy by content-based image retrieval', Retina, vol. 28, no. 10, pp. 1463-1477. https://doi.org/10.1097/IAE.0b013e31818356dd
Chaum E, Karnowski TP, Govindasamy VP, Abdelrahman M, Tobin KW. Automated diagnosis of retinopathy by content-based image retrieval. Retina. 2008 Nov 1;28(10):1463-1477. https://doi.org/10.1097/IAE.0b013e31818356dd
Chaum, Edward ; Karnowski, Thomas P. ; Govindasamy, V. Priya ; Abdelrahman, Mohamed ; Tobin, Kenneth W. / Automated diagnosis of retinopathy by content-based image retrieval. In: Retina. 2008 ; Vol. 28, No. 10. pp. 1463-1477.
@article{0b86d7994cd74f0bbd04813b364b51b9,
title = "Automated diagnosis of retinopathy by content-based image retrieval",
abstract = "PURPOSE: To describe a novel computer-based image analysis method that is being developed to assist and automate the diagnosis of retinal disease. METHODS: Content-based image retrieval is the process of retrieving related images from large database collections using their pictorial content. The content feature list becomes the index for storage, search, and retrieval of related images from a library based upon specific visual characteristics. Low-level analyses use feature description models and higher-level analyses use perceptual organization and spatial relationships, including clinical metadata, to extract semantic information. RESULTS: We defined, extracted, and tested a large number of region- and lesion-based features from a dataset of 395 retinal images. Using a statistical hold-one-out method, independent queries for each image were submitted to the system and a diagnostic prediction was formulated. The diagnostic sensitivity for all stratified levels of age-related macular degeneration ranged from 75{\%} to 100{\%}. Similarly, the sensitivity of detection and accuracy for proliferative diabetic retinopathy ranged from 75{\%} to 91.7{\%} and for nonproliferative diabetic retinopathy, ranged from 75{\%} to 94.7{\%}. The overall purity of the diagnosis (specificity) for all disease states in the dataset was 91.3{\%}. CONCLUSIONS: The probabilistic nature of content-based image retrieval permits us to make statistically relevant predictions regarding the presence, severity, and manifestations of common retinal diseases from digital images in an automated and deterministic manner.",
author = "Edward Chaum and Karnowski, {Thomas P.} and Govindasamy, {V. Priya} and Mohamed Abdelrahman and Tobin, {Kenneth W.}",
year = "2008",
month = "11",
day = "1",
doi = "10.1097/IAE.0b013e31818356dd",
language = "English (US)",
volume = "28",
pages = "1463--1477",
journal = "Retina",
issn = "0275-004X",
publisher = "Lippincott Williams and Wilkins",
number = "10",

}

TY - JOUR

T1 - Automated diagnosis of retinopathy by content-based image retrieval

AU - Chaum, Edward

AU - Karnowski, Thomas P.

AU - Govindasamy, V. Priya

AU - Abdelrahman, Mohamed

AU - Tobin, Kenneth W.

PY - 2008/11/1

Y1 - 2008/11/1

N2 - PURPOSE: To describe a novel computer-based image analysis method that is being developed to assist and automate the diagnosis of retinal disease. METHODS: Content-based image retrieval is the process of retrieving related images from large database collections using their pictorial content. The content feature list becomes the index for storage, search, and retrieval of related images from a library based upon specific visual characteristics. Low-level analyses use feature description models and higher-level analyses use perceptual organization and spatial relationships, including clinical metadata, to extract semantic information. RESULTS: We defined, extracted, and tested a large number of region- and lesion-based features from a dataset of 395 retinal images. Using a statistical hold-one-out method, independent queries for each image were submitted to the system and a diagnostic prediction was formulated. The diagnostic sensitivity for all stratified levels of age-related macular degeneration ranged from 75% to 100%. Similarly, the sensitivity of detection and accuracy for proliferative diabetic retinopathy ranged from 75% to 91.7% and for nonproliferative diabetic retinopathy, ranged from 75% to 94.7%. The overall purity of the diagnosis (specificity) for all disease states in the dataset was 91.3%. CONCLUSIONS: The probabilistic nature of content-based image retrieval permits us to make statistically relevant predictions regarding the presence, severity, and manifestations of common retinal diseases from digital images in an automated and deterministic manner.

AB - PURPOSE: To describe a novel computer-based image analysis method that is being developed to assist and automate the diagnosis of retinal disease. METHODS: Content-based image retrieval is the process of retrieving related images from large database collections using their pictorial content. The content feature list becomes the index for storage, search, and retrieval of related images from a library based upon specific visual characteristics. Low-level analyses use feature description models and higher-level analyses use perceptual organization and spatial relationships, including clinical metadata, to extract semantic information. RESULTS: We defined, extracted, and tested a large number of region- and lesion-based features from a dataset of 395 retinal images. Using a statistical hold-one-out method, independent queries for each image were submitted to the system and a diagnostic prediction was formulated. The diagnostic sensitivity for all stratified levels of age-related macular degeneration ranged from 75% to 100%. Similarly, the sensitivity of detection and accuracy for proliferative diabetic retinopathy ranged from 75% to 91.7% and for nonproliferative diabetic retinopathy, ranged from 75% to 94.7%. The overall purity of the diagnosis (specificity) for all disease states in the dataset was 91.3%. CONCLUSIONS: The probabilistic nature of content-based image retrieval permits us to make statistically relevant predictions regarding the presence, severity, and manifestations of common retinal diseases from digital images in an automated and deterministic manner.

UR - http://www.scopus.com/inward/record.url?scp=58949091123&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=58949091123&partnerID=8YFLogxK

U2 - 10.1097/IAE.0b013e31818356dd

DO - 10.1097/IAE.0b013e31818356dd

M3 - Article

VL - 28

SP - 1463

EP - 1477

JO - Retina

JF - Retina

SN - 0275-004X

IS - 10

ER -