List your company

Object Detection on VisDrone (test-dev)

61.3mAP@0.5Full Fine-tuning
3.16418.25733.3548.443Feb 5, 2026Mar 1, 2026Mar 25, 2026Apr 19, 2026May 13, 2026Jun 6, 2026Jul 1, 2026

Evaluation results

MethodLinks
Full Fine-tuning∆ Params.=0, ∆ Trainable.=27.5M2026.0761.3———38.4———————
DroneFINE-T (Ours)∆ Params.=1.54M, ∆ Trainable.=1.54M2026.0761.2———38.4———————
Mona∆ Params.=1.4M, ∆ Trainable.=1.4M2026.0760.5———37.9———————
Adapter∆ Params.=1.14M, ∆ Trainable.=1.14M2026.0760.1———37.4———————
Shine∆ Params.=0.02M, ∆ Trainable.=0.02M2026.0760.1———37———————
CoOp∆ Params.=0.012M, ∆ Trainable.=0.012M2026.0760———37.3———————
Baseline∆ Params.=0, ∆ Trainable.=02026.0759.6———37.1———————
NormTuning∆ Params.=0, ∆ Trainable.=0.025M2026.0759.6———37.1———————
VPT∆ Params.=0.08M, ∆ Trainable.=0.08M2026.0759.6———37.1———————
LoRA∆ Params.=1.13M, ∆ Trainable.=1.13M2026.0759.3———36.7———————
Proposed ModelVision Backbone=Proposed YOLOv8, Text Model=CLIP, GFLOPs=15.2, Params (M)=3.82026.0230.741.63135.5————————
YOLO-WorldVision Backbone=YOLOv8, Text Model=CLIP, GFLOPs=15.7, Params (M)=42026.0230.440.630.835————————
YOLOv9Vision Backbone=YOLOv9, GFLOPs=7.9, Params (M)=22026.0230.242.330.235.2————————
YOLOv11Vision Backbone=YOLOv11, GFLOPs=6.3, Params (M)=2.52026.0229.840.230.634.7————————
YOLOv10Vision Backbone=YOLOv10, GFLOPs=8.2, Params (M)=2.62026.0229.740.430.334.6————————
Zero-shot Detection YOLOVision Backbone=YOLOv5, Text Model=CLIP, GFLOPs=18.5, Params (M)=7.32026.0229.139.329.933.9————————
Zero-shot∆ Params.=0, ∆ Trainable.=–2026.0726.9———16.5———————
DLWBackbone=EfficientDet2026.0212.1———6.51.11033.29.72.915.742.1
DLSBackbone=EfficientDet2026.0211.9———6.611033.39.72.815.840.8
BaselineBackbone=EfficientDet2026.0211.2———6.20.99.533.49.32.715.240.8
BWBackbone=EfficientDet2026.0211———6.10.99.135.79.12.514.745.4
IWBackbone=EfficientDet2026.0211———6.10.99.334.29.22.614.843.3
CLBackbone=EfficientDet2026.026.8———3.70.65.121.861.79.230.8
LBackbone=EfficientDet2026.025.4———3.10.43.919.44.71.46.324.3