Interactive Segmentation on FocalClick
3.72Params (M)SegFormerB0
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| SegFormerB0Backbone=SegFormerB0, Input Resolution=448x448, Hardware=GPU: NVIDIA RTX4090, CPU: Intel(R) Xeon(R) Silver2026.03 | 3.72 | 1.75 | 37 | |
| FreqFormerBackbone=FreqFormer, Input Resolution=448x448, Hardware=GPU: NVIDIA RTX4090, CPU: Intel(R) Xeon(R) Silver2026.03 | 9.2 | 2.67 | 46 | |
| HRNet18Backbone=HRNet18, Input Resolution=448x448, Hardware=GPU: NVIDIA RTX4090, CPU: Intel(R) Xeon(R) Silver2026.03 | 10.03 | 30.99 | 56 | |
| ResNet34Backbone=ResNet34, Input Resolution=448x448, Hardware=GPU: NVIDIA RTX4090, CPU: Intel(R) Xeon(R) Silver2026.03 | 23.47 | 113.6 | 34 | |
| HRNet32Backbone=HRNet32, Input Resolution=448x448, Hardware=GPU: NVIDIA RTX4090, CPU: Intel(R) Xeon(R) Silver2026.03 | 30.95 | 83.12 | 86 | |
| Resnet50Backbone=Resnet50, Input Resolution=448x448, Hardware=GPU: NVIDIA RTX4090, CPU: Intel(R) Xeon(R) Silver2026.03 | 40.36 | 78.82 | 331 | |
| SegFormerB3Backbone=SegFormerB3, Input Resolution=448x448, Hardware=GPU: NVIDIA RTX4090, CPU: Intel(R) Xeon(R) Silver2026.03 | 45.66 | 24.75 | 53 | |
| Swin-BaseBackbone=Swin-Base, Input Resolution=448x448, Hardware=GPU: NVIDIA RTX4090, CPU: Intel(R) Xeon(R) Silver2026.03 | 87.44 | 138.21 | 36 | |
| ViT-BaseBackbone=ViT-Base, Input Resolution=448x448, Hardware=GPU: NVIDIA RTX4090, CPU: Intel(R) Xeon(R) Silver2026.03 | 96.46 | 42.44 | 54 | |
| Swin-LargeBackbone=Swin-Large, Input Resolution=448x448, Hardware=GPU: NVIDIA RTX4090, CPU: Intel(R) Xeon(R) Silver2026.03 | 195.9 | 302.78 | 44 | |
| ViT-LargeBackbone=ViT-Large, Input Resolution=448x448, Hardware=GPU: NVIDIA RTX4090, CPU: Intel(R) Xeon(R) Silver2026.03 | 322.18 | 133.22 | 86 |