Table 3 Experiments comparison results of different models.

From: An instance segmentation network for discharging carbon traces inside oil-immersed transformers with boundary and detail features enhancement

Model

Parameters (M)

FLOPs (G)

Precision

mAP

YOLOv3tiny-seg

14.1

32.8

75.8%

60.6%

YOLOv5s-seg

9.7

38.1

82.6%

66.8%

YOLOv6s-seg

16.7

56.7

84.9%

66.1%

YOLOv7tiny-seg

6.9

43.9

83.0%

63.5%

YOLOv9s-seg

11.2

76.9

84.1%

71.7%

YOLOv11s-seg

10.1

35.5

82.1%

67.2%

YOLOv12s-seg

9.9

35.6

83.7%

67.1%

DeepLabV3+

25.60

54.77

82.9%

66.3%

UNet3+

26.97

197.89

82.5%

67.8%

SwinUNet

96.34

42.68

79.1%

64.2%

CDEP-CTSeg

30.9

35.4

89.6%

76.7%

Back to article page