Skip to main content

Table 2 Classification performance parameters

From: An easy method to differentiate retinal arteries from veins by spectral domain optical coherence tomography: retrospective, observational case series

Performance measure

Using FA + IR as ground truth

Using IR as ground truth

Unclassified vessel

Unclassified vessel

Included

Excluded

included

Excluded

Artery

Vein

Artery

Vein

Artery

Vein

Artery

Vein

Sensitivity

0.7014

0.8000

0.8177

0.9172

0.7929

0.8947

0.8758

0.9297

Specificity

0.9278

0.8436

0.9172

0.8177

0.9323

0.8876

0.9297

0.8758

Positive Predictive Value

0.9193

0.8136

0.9193

0.8136

0.9371

0.8623

0.9371

0.8623

Negative Predictive Value

0.7261

0.8318

0.8136

0.9193

0.7799

0.9146

0.8623

0.9371

Positive Likelihood Ratio

9.7120

5.1152

9.8751

5.0307

11.7173

7.9584

12.4561

7.4864

Negative Likelihood Ratio

0.3218

0.2371

0.1988

0.1013

0.2221

0.1186

0.1336

0.0803

Classification Accuracy

0.8056

0.8235

0.8639

0.8639

0.8543

0.8907

0.9004

0.9004

Classification Error Rate

0.1944

0.1765

0.1361

0.1361

0.1457

0.1093

0.0996

0.0996

False Positive Rate

0.0722

0.1564

0.0828

0.1823

0.0677

0.1124

0.0703

0.1242

Back to article page

BMC Ophthalmology

ISSN: 1471-2415

Contact us