Skip to main content
. 2024 Aug 20;15:1396568. doi: 10.3389/fpls.2024.1396568

Table 2.

Testing performance (mAP@[0.5:0.95]) comparison between the supervised and semi-supervised based on Faster RCNN and FCOS models and CottonWeedDet12 dataset.

Algorithms Supervision type Proportion of labeled data for training
5% 10% 20% 50% 100%
Faster-RCNN Supervised 45.02 61.18 68.29 75.97 80.47
Semi-supervised 53.08 70.21 75.15 78.83
FCOS Supervised 62.28 72.99 79.14 83.87 86.47
Semi-supervised 76.91 83.43 85.28 87.26