Fig 1.
LiNet backbone including neck of BiFPN, PANet and FPN.
Fig 2.
Backbone component of YOLOv5-LiNet.
Stem, ResNet, Shuffle_Block, SPPF and C3 incorporated into neck network.
Table 1.
List of trained models.
Fig 3.
Output for box validation loss.
Fig 4.
Output for segmentation validation loss.
Fig 5.
Output for box F1 score.
Fig 6.
Output for segmentation F1 score.
Fig 7.
Output for box mAP@0.5.
Fig 8.
Output for segmentation mAP@0.5.
Fig 9.
Image results of YOLOv4-tiny.
Fig 10.
Image results of YOLOv5-Efficientlite.
Fig 11.
Image results of YOLOv5-MobileNetv3.
Fig 12.
Image results of YOLOv5-GhostNet.
Fig 13.
Image results of YOLOv5-ShuffleNetv2.
Fig 14.
Image results of YOLOv5n.
Fig 15.
Image results of YOLOv5-LiNetFPN.
Fig 16.
Image results of YOLOv5-LiNetBiFPN.
Fig 17.
Image results of YOLOv5-LiNetC.
Fig 18.
Image results of YOLOv5-LiNet.
Table 2.
Compared summary performance of the tested models.