Research Article
Multiloss Function Based Deep Convolutional Neural Network for Segmentation of Retinal Vasculature into Arterioles and Venules
Table 3
Quantitative performance measures of AV classification.
| Dataset | ā | Global Accuracy | Mean Accuracy | Mean IoU | Weighted IoU | Mean BF Score |
| AV-Classification | Pixel Level | 0.9652 | 0.8228 | 0.7905 | 0.9457 | 0.8201 | Multi-level (Segment/Pixel) | 0.9703 | 0.8739 | 0.8506 | 0.9568 | 0.8313 |
| DRIVE | Pixel Level | 0.9492 | 0.6823 | 0.55904 | 0.9207 | 0.74312 | Multi-level (Segment/Pixel) | 0.9607 | 0.6913 | 0.5666 | 0.9378 | 0.75547 |
| AVRDB | Pixel Level | 0.9650 | 0.7009 | 0.4322 | 0.8562 | 0.70134 | Multi-level (Segment/Pixel) | 0.9813 | 0.7109 | 0.4375 | 0.8602 | 0.70216 |
|
|