Skip to main content

Table 2 Results (pixel level accuracy) on 8 layered U-Net architecture initialized using He Normal. a) Results on the validation slides in dataset S1. b) Results on the test slides in dataset S1

From: Semantic segmentation to identify bladder layers from H&E Images

  a b
Accuracy True Positive False Negative Accuracy True Positive False Negative
Background 0.99 94,424,339 54,224 0.99 119,213,206 48,864
Lamina Propria 0.99 28,939,686 177,916 0.97 52,727,211 1,696,663
Muscularis Propria 0.88 110,788,573 14,326,899 0.90 22,782,045 2,649,277
Mucosa 0.87 11,433,732 1,750,370 0.90 37,076,397 4,034,633
RBC 0.92 9,996,911 848,300 0.93 1,884,697 146,841
Cautery 0.62 6,424,900 3,927,745 0.28 6,701,447 16,835,934
Inflammation 0.94 41,268,732 2,686,066 0.52 9,082,821 8,235,484
Muscularis Mucosa 0.89 1,511,039 169,144 N/A N/A N/A