Table 12 Comparison of semantic segmentation results with different weights.

From: Construction of a multiscale feature fusion model for indoor scene recognition and semantic segmentation

Evaluation indicators

k1 = 0.5, k2 = 0.5

k1 = 0.5, k2 = 1

k1 = 1, k2 = 0.5

k1 = 1, k2 = 1

mIoU/%

55.6

58.4

55.4

62.9

OA/%

76.4

78.8

76.3

82.6

Precision/%

65.0

67.1

65.0

70.4

Recall/%

75.1

77.5

75.0

81.3

F1/%

69.7

72.9

69.6

75.4