Object Detection Efficiency on SUN RGB-D (Latency & Size)
476.5GFLOPSsimCrossTrans
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| simCrossTransBackbone Network=Swin-T2022.03 | 476.5 | 48 | 105 | 9.5 | |
| simCrossTransBackbone Network=ResNet-502022.03 | 472.1 | 44 | 70 | 14.3 | |
| FETNetModality=RGB-D2023.09 | 279.3 | — | — | — | |
| YOLOv8xModality=RGB2023.09 | 258.5 | — | — | — | |
| YOLOv12xModality=RGB2023.09 | 200.3 | — | — | — | |
| YOLOv11xModality=RGB2023.09 | 195.6 | — | — | — | |
| Cascade R-CNNModality=RGB2023.09 | 168.3 | — | — | — | |
| Cascade R-CNN + FEM + MViTModality=RGB-D2023.09 | 158.5 | — | — | — | |
| Faster R-CNNModality=RGB2023.09 | 140.5 | — | — | — | |
| Faster R-CNN + FEM + MViTModality=RGB-D2023.09 | 130.7 | — | — | — | |
| YOLOv2Modality=RGB2023.09 | 63.03 | — | — | — | |
| Proposed ModelModality=RGB-D2023.09 | 26.72 | — | — | — | |
| F-VoxNetBackbone Network=ResNet-1012022.03 | — | 64 | 110 | 9.1 |