Supplementary Material: Density-aware Chamfer Distance as a Comprehensive Metric for Point Cloud Completion Tong Wu

We use Eqn. 1 in training for the loss between the coarse shape with 1024 points and the ground truth CD and the uniform loss are not sensitive to it. DCD and EMD can reflect the increasingly mismatched density, while CD and uniform loss are not sensitive to it. We introduce two hyper-parameters in Sec. Visualization of the distance contributed by each point. We invite 15 volunteers to complete the study.