Skip to main content

Table 1 Ablation experiments of different feature extraction modules

From: WheatLFANet: in-field detection and counting of wheat heads with high-real-time global regression network

Method

Feature Extractor

Pr

Re

AP

FPS

Latency

Params

FLOPs

WheatLFANet

RepVGG

0.896

0.835

0.887

157

8.6 ms

1.04 M

5.65G

ConvNeXt

0.878

0.827

0.859

90

13.3 ms

0.92 M

5.03G

ShuffleNetV2

0.881

0.802

0.859

170

8.3 ms

0.29 M

1.65G

CSPDarknet

0.905

0.843

0.900

164

8.4 ms

0.72 M

4.07G

  1. Bold text indicates the best results
Back to article page

Plant Methods

ISSN: 1746-4811