Skip to main content

Table 2 Pose prediction results of the 7 fixed views

From: Weakly-Supervised Single-view Dense 3D Point Cloud Reconstruction via Differentiable Renderer

Views Poses (quaternion) Error (\(^\circ\))
View 1 Ours − 0.310 0.181 0.419 0.833 1.046
GT 0.339 − 0.175 − 0.424 − 0.820
View 2 Ours − 0.396 0.828 0.347 0.188 2.871
GT 0.424 − 0.820 − 0.339 − 0.175
View 4 Ours − 0.173 0.339 − 0.822 − 0.421 0.292
GT 0.175 − 0.339 0.820 0.424
View 5 Ours − 0.149 0.333 0.821 0.438 1.354
GT 0.175 − 0.339 − 0.820 − 0.424
View 6 Ours − 0.815 0.432 0.174 0.342 0.995
GT 0.820 − 0.424 − 0.175 − 0.339
View 7 Ours 0.341 − 0.172 0.422 0.821 0.395
GT 0.339 − 0.175 0.424 0.820
View 8 Ours 0.416 − 0.823 0.344 0.174 0.996
GT 0.424 − 0.820 0.339 0.175