Table 3 Experimental data results.

From: A novel deep learning model based on YOLOv5 optimal method for coal gangue image recognition

  

P

R

mAP

Par

FLOPs

Times

YOLOv5 basic

Coal

0.963

0.949

0.973

–

–

–

Rock

0.963

0.959

0.977

–

–

–

All

0.963

0.954

0.975

7.03 Mb

16.0 G

6.9 ms

YOLOv5-CARAFE

Coal

0.951

0.961

0.972

–

–

–

Rock

0.961

0.966

0.979

–

–

–

All

0.956

0.956

0.976

7.16 Mb

16.4 G

8.0 ms

YOLOv5-MCA

Coal

0.965

0.947

0.974

–

–

–

Rock

0.959

0.959

0.979

–

–

–

All

0.971

0.936

0.976

7.06 Mb

16.1 G

7.7 ms

YOLOv5 optimal

Coal

0.962

0.959

0.98

–

–

–

Rock

0.971

0.96

0.975

–

–

–

All

0.966

0.959

0.977

7.19 Mb

16.6 G

9.0 ms

Back to article page