From: Prediction of broken areas in murals based on MLP-fused long-range semantics
Model | Params (M) | Flops (GMac) | Broken.IoU[%] | MIoU[%] | Dice coefficient[%] |
---|---|---|---|---|---|
GSCNN | 137.28 | 188.92 | 73.1 | 84.5 | 81.4 |
UNet +  +  | 47.19 | 200.13 | 68.9 | 82.0 | 79.0 |
DANet | 46.72 | 125.23 | 61.3 | 77.6 | 76.0 |
DPT | 124.0 | 51.39 | 67.6 | 81.3 | 80.7 |
Ours | 45.0 | 187.33 | 78.3 | 87.5 | 85.7 |