Table 8 Comparison with previous works based on the quantity of trainable parameters (Params), Gflops, and FPS.
From: A simple monocular depth estimation network for balancing complexity and accuracy
Method | NYU | KITTI | ||||||
|---|---|---|---|---|---|---|---|---|
Resolution | Gflops\(\downarrow\) | FPS\(\uparrow\) | Params\(\downarrow\) | Resolution | Gflops\(\downarrow\) | FPS\(\uparrow\) | Params\(\downarrow\) | |
AdaBins12 | \(416 \times 544\) | 137.18 | 28.4 | 78.3M | \(352 \times 704\) | 150.21 | 27.6 | 78.3M |
NewCRFs19 | \(480 \times 640\) | 280.46 | 18.4 | 270.4M | \(352 \times 1120\) | 361.13 | 14.7 | 270.4M |
LifelongDepth20 | \(480 \times 640\) | 63.29 | 108.1 | 22.3M | \(352 \times 1120\) | 81.22 | 98.1 | 22.3M |
IEBins18 | \(480 \times 640\) | 377.91 | 16.2 | 90.7M | \(352 \times 1120\) | 484.52 | 13.8 | 90.7M |
Metric3Dv229 | \(480 \times 640\) | 263.04 | 20.6 | 203.24M | \(352 \times 1120\) | 337.57 | 16.2 | 203.24M |
SimMDE(Ours) | \(480 \times 640\) | 42.79 | 28.8 | 30.9M | \(352 \times 704\) | 46.85 | 30.8 | 30.9M |
