Table 2 Detailed classification performance of LightMask segmented by class.

From: A lightweight transformer framework for open set anomaly segmentation in smart city applications

Class

Precision

Recall

F1-Score

Support

Normal

0.99

0.70

0.82

2,731,954

Anomaly

0.25

0.93

0.40

296,606

Accuracy

–

–

0.72

3,010,560

Macro Average

0.62

0.81

0.61

3,010,560

Weighted Average

0.92

0.72

0.78

3,010,560