Table 1 Results of accuracy metrics and the number of convolutional layers, trainable parameters, training epochs, total training time, giga floating point operations per second (GFLOPS), and processing time per image for all trained models.

From: Automated assessment of periapical health based on the radiographic periapical index using YOLOv8, YOLOv11, and YOLOv12 one-stage object detection algorithms

 

YOLOv8m

YOLOv11m

YOLOv12m

layers

92

125

169

Trainable parameters

25,842,655

20,033,887

20,108,767

Training epochs

421

521

399

Training time (hours)

12.140

9.870

10.350

GFLOPs

78.7

67.7

67.1

Processing time (ms)

3

5.4

6.8

mAP50

86.4%

86.6%

86.6%

Recall

83.5%

86.2%

84.4%

Precision

86.8%

88.5%

89.1%

Max F1 score (%)

85.2%

87.1%

86.6%

Back to article page