Semantic Segmentation on AOI Dataset
60.3mIoUFasterViT-0 + UPerNet
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| FasterViT-0 + UPerNetPre-training=MAE, Throughput (Crops/s)=163.32026.05 | 60.3 | 79.3 | 66.7 | 19.3 | 75.8 | |
| FasterViT-0 + UPerNetPre-training=iBOT2026.05 | 53.7 | 67.5 | 60 | 19.7 | 67.7 | |
| ViT-Tiny + UPerNetPre-training=MAE, Throughput (Crops/s)=218.12026.05 | 53.5 | 73.7 | 43 | 26.5 | 70.7 | |
| ResNet18 + U-Net++Pre-training=ImageNet, Throughput (Crops/s)=86.92026.05 | 52.4 | 75.4 | 59.1 | 9.2 | 65.8 | |
| ViT-T + Patch RetrievalPre-training=MAE, Throughput (Crops/s)=266.72026.05 | 48.1 | 38.4 | 47.8 | 44.4 | 61.8 | |
| ViT-S Patch RetrievalPre-training=DINOv2, Throughput (Crops/s)=90.92026.05 | 47.8 | 43.4 | 48.7 | 45.8 | 53.4 | |
| ViT-T + Patch RetrievalPre-training=iBOT, Throughput (Crops/s)=266.72026.05 | 45.9 | 46.2 | 43.7 | 41 | 52.8 | |
| ResNet18 + DeepLabPre-training=ImageNet, Throughput (Crops/s)=87.32026.05 | 43.5 | 66.7 | 50.3 | 10.3 | 46.7 | |
| ViT-Tiny + UPerNetPre-training=iBOT, Throughput (Crops/s)=218.12026.05 | 41.6 | 71.8 | 39.3 | 5.9 | 49.5 | |
| ViT-T + Patch RetrievalPre-training=DINO2026.05 | 41.5 | 41.5 | 41.1 | 36.6 | 46.9 | |
| ViT-Tiny + UPerNetPre-training=DINO2026.05 | 40 | 68.7 | 31.6 | 1.3 | 54.2 | |
| MobileNetV3 + DeepLabThroughput (Crops/s)=309.82026.05 | 29.9 | 60.6 | 19.3 | 2.8 | 36.8 |