About the function problem of converting point set to pseudo box

[davidyang180]

Hello! Through experiments, we will use the function method of moment to visualize the point set on some visual tasks and find that the distribution of the point set tends to cover the vicinity of the bounding box, while the actual points on the object are relatively few;
This phenomenon is intuitively inconsistent. Intuitively, the point set should be uniformly covered on the object. Is there a certain defect in the method of converting this point set into a bounding box?