Skip to main content
. 2025 Jan 7;15:1508549. doi: 10.3389/fpls.2024.1508549

Table 9.

Performance comparison based on the Banana Plantation dataset.

Model Backbone mIoU Params (M) FPS
DeepLabV3plus (Chen et al., 2018) R-101-D8 0.9283 60.21 6.98
KNet (Zhang et al., 2021) Swin-L 0.9468 245 4.55
Mask2Former (Cheng et al., 2022) Swin-L 0.9427 235 5.09
PSPNet (Zhao et al., 2017) R-50-D8 0.9113 46.60 11.91
Segformer (Xie et al., 2021) MIT-B5 0.9398 81.97 6.57
UNet (Ronneberger et al., 2015) UNet-S5-D16 0.9031 28.99 17.93
ALSS-YOLO-Seg ALSSNet 0.9272 1.8256 130.6