Automated detection and differentiation of drusen, exudates, and cotton-wool spots in digital color fundus photographs for diabetic retinopathy diagnosis

被引:265
作者
Niemeijer, Meindert
van Ginneken, Bram
Russell, Stephen R.
Suttorp-Schulten, Maria S. A.
Abramoff, Michael D.
机构
[1] Univ Iowa Hosp & Clin, Retina Serv, Dept Ophthalmol & Visual Sci, Iowa City, IA 52242 USA
[2] Univ Utrecht, Med Ctr, Image Sci Inst, Utrecht, Netherlands
[3] Iowa City VA Med Ctr, Dept Vet Affairs, Iowa City, IA USA
[4] OLVG, Ophthalmol Serv, Amsterdam, Netherlands
[5] Univ Iowa, Dept Elect & Comp Engn, Iowa City, IA USA
关键词
D O I
10.1167/iovs.06-0996
中图分类号
R77 [眼科学];
学科分类号
100212 ;
摘要
PURPOSE. To describe and evaluate a machine learning-based, automated system to detect exudates and cotton-wool spots in digital color fundus photographs and differentiate them from drusen, for early diagnosis of diabetic retinopathy. METHODS. Three hundred retinal images from one eye of 300 patients with diabetes were selected from a diabetic retinopathy telediagnosis database (nomydriatic camera, two-field photography): 100 with previously diagnosed bright lesions and 200 without. A machine learning computer program was developed that can identify and differentiate among drusen, (hard) exudates, and cotton-wool spots. A human expert standard for the 300 images was obtained by consensus annotation by two retinal specialists. Sensitivities and specificities of the annotations on the 300 images by the automated system and a third retinal specialist were determined. RESULTS. The system achieved an area under the receiver operating characteristic (ROC) curve of 0.95 and sensitivity/specificity pairs of 0.95/0.88 for the detection of bright lesions of any type, and 0.95/0.86, 0.70/0.93, and 0.77/0.88 for the detection of exudates, cotton-wool spots, and drusen, respectively. The third retinal specialist achieved pairs of 0.95/0.74 for bright lesions and 0.90/0.98, 0.87/0.98, and 0.92/0.79 per lesion type. CONCLUSIONS. A machine learning-based, automated system capable of detecting exudates and cotton-wool spots and differentiating them from drusen in color images obtained in community based diabetic patients has been developed and approaches the performance level of retinal experts. If the machine learning can be improved with additional training data sets, it may be useful for detecting clinically important bright lesions, enhancing early diagnosis, and reducing visual loss in patients with diabetes.
引用
收藏
页码:2260 / 2267
页数:8
相关论文
共 30 条
[1]   Web-based screening for diabetic retinopathy in a primary care population: The EyeCheck project [J].
Abramoff, MD ;
Suttorp-Schulten, MSA .
TELEMEDICINE JOURNAL AND E-HEALTH, 2005, 11 (06) :668-674
[2]  
Abramoff MD, 2006, 28 ANN INT C IEEE EN, P4432, DOI DOI 10.1109/IEMBS.2006.259622
[3]   Automated segmentation of the optic disc from stereo color photographs using physiologically plausible features [J].
Abramoff, Michael D. ;
Alward, Wallace L. M. ;
Greenlee, Emily C. ;
Shuba, Lesya ;
Kim, Chan Y. ;
Fingert, John H. ;
Kwon, Young H. .
INVESTIGATIVE OPHTHALMOLOGY & VISUAL SCIENCE, 2007, 48 (04) :1665-1673
[4]   A screening approach to the surveillance of patients with diabetes for the presence of vision-threatening retinopathy [J].
Bresnick, GH ;
Mukamel, DB ;
Dickinson, JC ;
Cole, DR .
OPHTHALMOLOGY, 2000, 107 (01) :19-24
[5]  
Duda RO, 2006, PATTERN CLASSIFICATI
[6]  
Early Treatment Diabetic Retinopathy Study Res Grp, 1991, OPHTHALMOLOGY, V98, P766
[7]   Detection of optic disc in retinal images by means of a geometrical model of vessel structure [J].
Foracchia, M ;
Grisan, E ;
Ruggeri, A .
IEEE TRANSACTIONS ON MEDICAL IMAGING, 2004, 23 (10) :1189-1195
[8]   Automatic detection of diabetic retinopathy using an artificial neural network: A screening tool [J].
Gardner, GG ;
Keating, D ;
Williamson, TH ;
Elliott, AT .
BRITISH JOURNAL OF OPHTHALMOLOGY, 1996, 80 (11) :940-944
[9]  
Grading diabetic retinopathy from stereoscopic color fundus photographsean extension of the modified Airlie House classification, 1991, OPHTHALMOLOGY S, V98, P786
[10]   Locating blood vessels in retinal images by piecewise threshold probing of a matched filter response [J].
Hoover, A ;
Kouznetsova, V ;
Goldbaum, M .
IEEE TRANSACTIONS ON MEDICAL IMAGING, 2000, 19 (03) :203-210