Figure 5
From: Scoring pleurisy in slaughtered pigs using convolutional neural networks
Scoring through a rule-based classifier: an example. The rule-based classifier selected 3 layers of the segmented picture, namely caudal chest wall (A), cranial chest wall (B) and pleural lesions (C). Thereafter, the same classifier checked the overlapping among such districts, in order to assign the score: (D) the pleural lesion (red color) partially overlaps with the cranial chest wall (dark blue color); (E) the pleural lesion (red color) partially overlaps with the caudal chest wall (light blue color). Total score = 3.