Research Article
Retina Image Vessel Segmentation Using a Hybrid CGLI Level Set Method
Table 2
The performance of our CGLI method on the DRIVE data and STARE data compared with other methods. The result of our method is in italic.
| | Method | DRIVE | STARE | sn | sp | acc | sn | sp | acc |
| Unsupervised method | CGLI | 0.7358 | 0.9680 | 0.9390 | 0.7449 | 0.9690 | 0.9409 | Azzopardi et al. (2015) [1] | 0.7655 | 0.9704 | 0.9442 | 0.7716 | 0.9701 | 0.9497 | Vlachos and Dermatas (2010) [8] | 0.7468 | 0.9551 | 0.9285 | 0.7455 | 0.9544 | 0.9270 | Martí et al. (2007) [9] | 0.6634 | 0.9682 | 0.9352 | 0.6701 | 0.9599 | 0.9371 |
| Supervised method | Marín et al. (2011) [10] | NA | NA | 0.9452 | NA | NA | 0.9344 | Fraz et al. (2012) [11] | 0.7152 | 0.9759 | 0.9430 | 0.7311 | 0.9680 | 0.9442 |
|
|