BMC Medical Imaging

Table 2 Binary classification results after the three groups are collapsed into binary groups

From: Use of multimodal dataset in AI for detecting glaucoma based on fundus photographs assessed with OCT: focus group study on high prevalence of myopia

Metrics	Validation dataset		Test dataset
Metrics	(N^a, PPG^b + G^c)	(N^a + PPG^b, G^c)	(N^a, PPG^b + G^c)	(N^a + PPG^b, G^c)
AUROC^d, %	95.91	96.42^e	95.45^f	95.24
Accuracy, %	89.96	90.57^e	88.88^f	86.11
Sensitivity, %	89.93	90.54^e	88.63^f	86.77
Specificity, %	89.99	90.59^e	89.17^f	85.49
Precision, %	87.92	94.13^e	96.53^f	95.46
F3^g, %	89.73	89.80^e	88.77^f	86.56

^aN: normal
^bPPG: pre-perimetrical glaucoma
^cG: glaucoma
^dAUROC: area under the receiver operator characteristic curve
^ebest metrics value in the validation set
^fbest metrics value in the test set
^gF3: F-Beta measure, beta = 3

Back to article page

ISSN: 1471-2342

Contact us

General enquiries: journalsubmissions@springernature.com