Table 4 Comparison of mAP and Inference Speed (FPS) for YOLO versions.

From: Advanced gesture recognition in Indian sign language using a synergistic combination of YOLOv10 with Swin Transformer model

Model

mAP (Image)

mAP (Video)

FPS (Image)

FPS (Video)

YOLOv3

91.34%

89.76%

32.3

40.8

YOLOv4

92.12%

90.03%

34.7

32.5

YOLOv5

92.48%

90.87%

36.1

33.9

YOLOv6

92.25%

91.12%

36.4

34.2

YOLOv7

93.32%

92.18%

38.7

35.9

YOLOv8

94.21%

92.48%

39.2

36.7

YOLOv9

95.43%

93.97%

41.5

39.1

YOLOv10

96.80%

94.87%

44.6

41.3

YOLOv10-ST (Proposed)

97.62%

95.94%

48.7

45.5

Back to article page