Table 6 Ablation study on RM. Both global and local features play a crucial role in refinement. Identical weights of the global feature extractor (GFE) are utilized for both \({PC}_t\) and \({PC}_{t+1}\).
From: Scene flow based deep network for hand reconstruction using depth images
Sr. | Features | MPVPE (mm) |
|---|---|---|
1 | \({GF}_t\) only | 10.97 |
2 | \({LF}_{t}\) only | 10.75 |
3 | \({GF}_t\) only with shared GFE | 10.75 |
4 | Both \(\varvec{{GF}_t}\) and \(\varvec{{LF}_{t}}\) with shared \(\varvec{{GFE}}\) | 10.50 |
