Object Detection on Pascal VOC (mAP)
82.4mAPAutoMix
Evaluation results
| Method | Links | |
|---|---|---|
| AutoMixDetector=Faster-RCNN2021.03 | 82.4 | |
| ResizeMixDetector=Faster-RCNN2021.03 | 82.1 | |
| CutMixDetector=Faster-RCNN2021.03 | 81.9 | |
| PuzzleMixDetector=Faster-RCNN2021.03 | 81.9 | |
| VanillaDetector=Faster-RCNN2021.03 | 81 | |
| MixupDetector=Faster-RCNN2021.03 | 80.7 | |
| OVORImg Enc=CLIP, Training protocol=training-free2026.03 | 72.6 | |
| ViLDTraining protocol=Extra training2026.03 | 72.2 | |
| DK-DETRTraining protocol=Extra training2026.03 | 71.3 | |
| FasterRCNNBit-width=6 bit, Quantisation=Linear, Fine-tuning=Evolutionary2026.04 | 70.5 | |
| LLMDet (Tiny) + DetRefinerDetRefiner=Applied (trained on OV-LVIS), Trained Data=GroundingCap-1M2026.05 | 61.3 | |
| MM-Grounding DINO (Tiny) + DetRefinerDetRefiner=Applied (trained on OV-LVIS), Trained Data=O365, GoldG, GRIT, V3Det2026.05 | 60.4 | |
| Grounding DINO (Tiny) + DetRefinerDetRefiner=Applied (trained on OV-LVIS), Trained Data=O365, GoldG, Cap4M2026.05 | 60 | |
| LLMDet (Tiny)DetRefiner=None, Trained Data=GroundingCap-1M2026.05 | 59.4 | |
| GLIP (Tiny) + DetRefinerDetRefiner=Applied (trained on OV-LVIS), Trained Data=O365, GoldG, CC3M, SBU2026.05 | 59.2 | |
| MM-Grounding DINO (Tiny)DetRefiner=None, Trained Data=O365, GoldG, GRIT, V3Det2026.05 | 57.6 | |
| GLIP (Tiny)DetRefiner=None, Trained Data=O365, GoldG, CC3M, SBU2026.05 | 56.6 | |
| Grounding DINO (Tiny)DetRefiner=None, Trained Data=O365, GoldG, Cap4M2026.05 | 56.3 | |
| OVORImg Enc=MLP, Training protocol=Extra training2026.03 | 39.5 |