Research Article
Automatic Detection of Blood Vessels in Retinal Images for Diabetic Retinopathy Diagnosis
Table 5
Performance comparisons on DRIVE dataset.
| Algorithm | Sensitivity (%) | Specificity (%) | Accuracy (%) |
| Proposed method | 93.99 | 98.37 | 98.08 | Fraz et al. [7] (2012) | 74.06 | 98.07 | 94.8 | Xiao et al. [8] (2013) | 75.13 | 97.92 | 95.29 | Manoj et al. [9] (2013) | 94.29 | 98.75 | 96.23 | Bansal and Dutta [10] (2013) | 86.53 | 98.33 | 97.28 | Marín et al. [5] (2011) | 70.67 | 98.01 | 94.52 | Mendonça and Campilho [6] (2006) | ⋯ | ⋯ | 96.33 | Palomera-Pérez et al. [4] (2010) | ⋯ | ⋯ | 92.5 | Budai et al. [11] (2013) | 64.40 | 98.70 | 95.72 |
|
|
The methodologies compared with proposed method did not consider OD elimination before vessel segmentation.
|