Table 1 Stage 1: Improved model ablation experiments.

From: Research on pedestrian recognition in complex scenarios based on data augmentation using large language models

Models

Params/M

GFLOPs

P/%

R/%

mAP@0.5/%

mAP@0.5:0.95/%

YOLOv8n (baseline)

11.17

28.8

86.7

80.9

89.5

63.6

 + C2f-RVB

8.25

21.5

86.8

81.2

89.6

63.9

 + C2f-RVE

8.27

21.7

87.9

82.0

90.0

64.6

 + C2f-RVB-GSConv

7.89

21.0

87.1

81.2

89.8

63.9

 + C2f-RVE-GSConv (Ours)

7.91

21.2

87.8

81.8

90.0

64.4

Back to article page