Screening fundus images for diabetic retinopathy

Sohini Roychowdhury, Dara Koozekanani, Keshab K Parhi

Research output: Chapter in Book/Report/Conference proceedingConference contribution

12 Citations (Scopus)

Abstract

This paper presents a novel two-stage system that detects diabetic retinopathy (DR) using fundus photographs. The first-stage of this system masks out the background consisting of the optic disc, using a novel Minimum Intensity Maximum Solidity (MinIMaS) overlap algorithm that is 99.7% accurate in segmenting the optic disc region on public data sets named DIARETDB0, DIARETDB1, DRIVE, and STARE. Receiver operating characteristic (ROC) analysis on the DIARETDB1 data set depicts that the second-stage of the system classifies bright lesions with 82.87% sensitivity, 94.36% specificity, 0.9593 area under ROC curves (AUC), and it detects red lesions with 75.5% sensitivity, 93.73% specificity, 0.8663 AUC using the Gaussian Mixture Models. Also, for DIARETDB1, free-response receiver operating characteristic (FROC) analysis shows that the proposed detection system achieves a sensitivity of 80% for bright lesion detection, and 64% for red lesion detection at 0.5 false positives per image. Thus, the proposed DR detection system outperforms existing works by lowering false positives in lesion classification, and hence it can be applied to enhance the effectiveness in screening patients for diabetic retinopathy.

Original languageEnglish (US)
Title of host publicationConference Record of the 46th Asilomar Conference on Signals, Systems and Computers, ASILOMAR 2012
Pages1641-1645
Number of pages5
DOIs
StatePublished - Dec 1 2012
Event46th Asilomar Conference on Signals, Systems and Computers, ASILOMAR 2012 - Pacific Grove, CA, United States
Duration: Nov 4 2012Nov 7 2012

Other

Other46th Asilomar Conference on Signals, Systems and Computers, ASILOMAR 2012
CountryUnited States
CityPacific Grove, CA
Period11/4/1211/7/12

Fingerprint

Optics
Screening
Masks

Keywords

  • classification
  • Diabetic retinopathy
  • exudates
  • optic disc
  • vascular arc

Cite this

Roychowdhury, S., Koozekanani, D., & Parhi, K. K. (2012). Screening fundus images for diabetic retinopathy. In Conference Record of the 46th Asilomar Conference on Signals, Systems and Computers, ASILOMAR 2012 (pp. 1641-1645). [6489310] https://doi.org/10.1109/ACSSC.2012.6489310

Screening fundus images for diabetic retinopathy. / Roychowdhury, Sohini; Koozekanani, Dara; Parhi, Keshab K.

Conference Record of the 46th Asilomar Conference on Signals, Systems and Computers, ASILOMAR 2012. 2012. p. 1641-1645 6489310.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Roychowdhury, S, Koozekanani, D & Parhi, KK 2012, Screening fundus images for diabetic retinopathy. in Conference Record of the 46th Asilomar Conference on Signals, Systems and Computers, ASILOMAR 2012., 6489310, pp. 1641-1645, 46th Asilomar Conference on Signals, Systems and Computers, ASILOMAR 2012, Pacific Grove, CA, United States, 11/4/12. https://doi.org/10.1109/ACSSC.2012.6489310
Roychowdhury S, Koozekanani D, Parhi KK. Screening fundus images for diabetic retinopathy. In Conference Record of the 46th Asilomar Conference on Signals, Systems and Computers, ASILOMAR 2012. 2012. p. 1641-1645. 6489310 https://doi.org/10.1109/ACSSC.2012.6489310
Roychowdhury, Sohini ; Koozekanani, Dara ; Parhi, Keshab K. / Screening fundus images for diabetic retinopathy. Conference Record of the 46th Asilomar Conference on Signals, Systems and Computers, ASILOMAR 2012. 2012. pp. 1641-1645
@inproceedings{fffe3fd03790424e90312ed19663d66b,
title = "Screening fundus images for diabetic retinopathy",
abstract = "This paper presents a novel two-stage system that detects diabetic retinopathy (DR) using fundus photographs. The first-stage of this system masks out the background consisting of the optic disc, using a novel Minimum Intensity Maximum Solidity (MinIMaS) overlap algorithm that is 99.7{\%} accurate in segmenting the optic disc region on public data sets named DIARETDB0, DIARETDB1, DRIVE, and STARE. Receiver operating characteristic (ROC) analysis on the DIARETDB1 data set depicts that the second-stage of the system classifies bright lesions with 82.87{\%} sensitivity, 94.36{\%} specificity, 0.9593 area under ROC curves (AUC), and it detects red lesions with 75.5{\%} sensitivity, 93.73{\%} specificity, 0.8663 AUC using the Gaussian Mixture Models. Also, for DIARETDB1, free-response receiver operating characteristic (FROC) analysis shows that the proposed detection system achieves a sensitivity of 80{\%} for bright lesion detection, and 64{\%} for red lesion detection at 0.5 false positives per image. Thus, the proposed DR detection system outperforms existing works by lowering false positives in lesion classification, and hence it can be applied to enhance the effectiveness in screening patients for diabetic retinopathy.",
keywords = "classification, Diabetic retinopathy, exudates, optic disc, vascular arc",
author = "Sohini Roychowdhury and Dara Koozekanani and Parhi, {Keshab K}",
year = "2012",
month = "12",
day = "1",
doi = "10.1109/ACSSC.2012.6489310",
language = "English (US)",
isbn = "9781467350518",
pages = "1641--1645",
booktitle = "Conference Record of the 46th Asilomar Conference on Signals, Systems and Computers, ASILOMAR 2012",

}

TY - GEN

T1 - Screening fundus images for diabetic retinopathy

AU - Roychowdhury, Sohini

AU - Koozekanani, Dara

AU - Parhi, Keshab K

PY - 2012/12/1

Y1 - 2012/12/1

N2 - This paper presents a novel two-stage system that detects diabetic retinopathy (DR) using fundus photographs. The first-stage of this system masks out the background consisting of the optic disc, using a novel Minimum Intensity Maximum Solidity (MinIMaS) overlap algorithm that is 99.7% accurate in segmenting the optic disc region on public data sets named DIARETDB0, DIARETDB1, DRIVE, and STARE. Receiver operating characteristic (ROC) analysis on the DIARETDB1 data set depicts that the second-stage of the system classifies bright lesions with 82.87% sensitivity, 94.36% specificity, 0.9593 area under ROC curves (AUC), and it detects red lesions with 75.5% sensitivity, 93.73% specificity, 0.8663 AUC using the Gaussian Mixture Models. Also, for DIARETDB1, free-response receiver operating characteristic (FROC) analysis shows that the proposed detection system achieves a sensitivity of 80% for bright lesion detection, and 64% for red lesion detection at 0.5 false positives per image. Thus, the proposed DR detection system outperforms existing works by lowering false positives in lesion classification, and hence it can be applied to enhance the effectiveness in screening patients for diabetic retinopathy.

AB - This paper presents a novel two-stage system that detects diabetic retinopathy (DR) using fundus photographs. The first-stage of this system masks out the background consisting of the optic disc, using a novel Minimum Intensity Maximum Solidity (MinIMaS) overlap algorithm that is 99.7% accurate in segmenting the optic disc region on public data sets named DIARETDB0, DIARETDB1, DRIVE, and STARE. Receiver operating characteristic (ROC) analysis on the DIARETDB1 data set depicts that the second-stage of the system classifies bright lesions with 82.87% sensitivity, 94.36% specificity, 0.9593 area under ROC curves (AUC), and it detects red lesions with 75.5% sensitivity, 93.73% specificity, 0.8663 AUC using the Gaussian Mixture Models. Also, for DIARETDB1, free-response receiver operating characteristic (FROC) analysis shows that the proposed detection system achieves a sensitivity of 80% for bright lesion detection, and 64% for red lesion detection at 0.5 false positives per image. Thus, the proposed DR detection system outperforms existing works by lowering false positives in lesion classification, and hence it can be applied to enhance the effectiveness in screening patients for diabetic retinopathy.

KW - classification

KW - Diabetic retinopathy

KW - exudates

KW - optic disc

KW - vascular arc

UR - http://www.scopus.com/inward/record.url?scp=84876246779&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84876246779&partnerID=8YFLogxK

U2 - 10.1109/ACSSC.2012.6489310

DO - 10.1109/ACSSC.2012.6489310

M3 - Conference contribution

SN - 9781467350518

SP - 1641

EP - 1645

BT - Conference Record of the 46th Asilomar Conference on Signals, Systems and Computers, ASILOMAR 2012

ER -