Research Article
Computer-Assisted Diagnosis for Diabetic Retinopathy Based on Fundus Images Using Deep Convolutional Neural Network
Table 7
Differences between Tables
5 and
6.
| Determined lesion | Ground-truth lesion | SVM-network 1 | SVM-network 2 | 0 | 1 | 2 | 3 | 4 | 0 | 1 | 2 | 3 | 4 |
| 0 | +420 | −460 | −848 | −31 | −80 | +310 | 0 | −104 | −43 | +2 | 1 | +194 | +393 | +164 | +1 | 0 | −15 | +96 | +59 | 0 | −2 | 2 | +231 | +70 | +648 | −37 | +1 | −220 | −90 | +197 | +81 | +48 | 3 | +3 | 0 | +45 | +76 | +21 | +1 | −1 | −123 | −78 | −62 | 4 | +3 | −3 | +45 | +76 | +58 | −65 | −5 | −28 | +4 | +14 | Accuracy increment (%) | −1.08 | 89.77 | 12.80 | 22.88 | 10.52 | 0.78 | 13.05 | 3.69 | −16.03 | 2.3 | Overall accuracy increment (%) | 1.6 | 1.16 |
|
|