Research Article
Automatic Detection of Small Intestinal Hookworms in Capsule Endoscopy Images Based on a Convolutional Neural Network
Table 2
Causes of discrepancies in classifications by the experts and by the CNN.
| | (%) |
| False-negative lesions () | | Poorly demarcated | (53.6) | Debris | 12 (21.4) | Darkness | 17 (30.4) | Poor focus | 1 (1.8) | Similarity to the edge of the bubble | 7 (12.5) | Similarity to the submucosal vascular shadow | 17 (30.4) | Smallness | 2 (3.6) | False-positive lesions () | | Darkness | 32 (3.6) | Bubble | 179 (20) | Debris | 516 (57.7) | Vascular shadow | 73 (8.2) | Smallness | 3 (0.2) | Fold | 92 (10.3) | Hookworms overlooked by experts () | | True hookworms | 2 (100) |
|
|
Caused by dark view or coffee-like bloody fluid. |