【3D deep learning on Point cloud】(1)

---

Point cloud generation

Set comparison: given two sets of points, measure their discrepancy

key challenge: correspondence problem

Correspondence(1): optimal assignment

　    Earth Mover's distance(EMD)

Correspondence(2):closest point

　　Chamfer distance(CD)：倒角距离。一种对于图像的距离变换，对于有特征点和非特征点的二值图像，此距离变换就是求解每一个点到最近特征点的距离。

　　参考：简书

　　         CSDN

---

Required properties of distance metrics

　　Geometric requirement

　　　　reflects natural shape differences

　　　　induce a nice space for shape interpolations

　　Computational requirement

　　　　define a loss function that is numerically easy to optimize

　　　　　to be used as a loss function, the metric has to be:

• • differentiable with respect to point locations
  • efficient to compute