Object Detection on FDTOOTH (test)
62.45AP75 (FD)FD-SOS
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| FD-SOSmulti-task=true, fine-tuning existing VLM=true, textual backbone=BERT-base, image backbone=SWIN-T, complexity=64.02M2024.07 | 62.45 | 60.84 | 66.01 | 67.07 | 65.97 | 69.67 | |
| SparseDetmulti-task=true, pre-training on publicly dental dataset=true, backbone=SWIN-T2024.07 | 60.5 | 52.94 | 62.95 | 62.96 | 54.87 | 65.13 | |
| GDINOmulti-task=false, fine-tuning existing VLM=true, textual backbone=BERT-base, image backbone=SWIN-T, complexity=64.02M2024.07 | 58.32 | 56.59 | 61.07 | 63.69 | 62.59 | 65.89 | |
| DDETRmulti-task=false, pre-training on publicly dental dataset=true, backbone=SWIN-T2024.07 | 56.92 | 50.41 | 60.51 | 62.68 | 57.44 | 65.48 | |
| Hierarchical-Diff-DetRmulti-task=true, pre-training on publicly dental dataset=true, backbone=SWIN-T2024.07 | 55.81 | 50.82 | 59.21 | 61.45 | 57.63 | 64.4 | |
| GDINO (our baseline)multi-task=true, fine-tuning existing VLM=true, textual backbone=BERT-base, image backbone=SWIN-T, complexity=64.02M2024.07 | 55.55 | 54.75 | 59.99 | 62.6 | 62.08 | 65.81 | |
| Diffusion-DETRmulti-task=false, pre-training on publicly dental dataset=true, backbone=SWIN-T2024.07 | 55.52 | 51.42 | 61.28 | 62.58 | 59.06 | 66.37 | |
| DINOmulti-task=false, pre-training on publicly dental dataset=true, backbone=SWIN-T, complexity=48.04M2024.07 | 54.03 | 49.68 | 57.94 | 55.13 | 51.65 | 57.65 | |
| DDETR (multi-task)multi-task=true, pre-training on publicly dental dataset=true, backbone=SWIN-T2024.07 | 53.28 | 47.57 | 58.01 | 60.2 | 54.89 | 63.05 | |
| DINO (multi-task)multi-task=true, pre-training on publicly dental dataset=true, backbone=SWIN-T, complexity=48.04M2024.07 | 48.12 | 43.85 | 50.69 | 53.8 | 50.15 | 55.53 | |
| GLIP (multi-task)multi-task=true, fine-tuning existing VLM=true, textual backbone=BERT-base, image backbone=SWIN-T, complexity=122.8M2024.07 | 41.78 | 33.68 | 47.09 | 51.97 | 42.73 | 56.7 | |
| GLIPmulti-task=false, fine-tuning existing VLM=true, textual backbone=BERT-base, image backbone=SWIN-T, complexity=122.8M2024.07 | 40.57 | 32 | 46.34 | 51.3 | 40.47 | 55.85 | |
| Diffusion-DETR w/o pretrainingmulti-task=false, pre-training on publicly dental dataset=false, backbone=SWIN-T2024.07 | 0.04 | 1.31 | 7.58 | 0.04 | 1.7 | 8.85 | |
| Hierarchical-Diff-DetR w/o pretrainingmulti-task=true, pre-training on publicly dental dataset=false, backbone=SWIN-T2024.07 | 0 | 0 | 0 | 0 | 0 | 0 |