Skip to main content
. 2023 Feb 6;9(2):35. doi: 10.3390/jimaging9020035

Figure 2.

Figure 2

Inference results on the UV dataset; each line contains (1) original images, (2) ground truth, (3) result from PVT_DA2 (i.e., the best stand-alone approach), and (4) AllM_H (the best ensemble). False-positive pixels are in green, while the false negatives are in red.