Google Inc.

Siemens Corporation

Lingyun Liu

Christopher Richard Uhlik

Tong Fang

Luc Vincent

Daniel Joseph Filip

Xiaoxu Ma

Hui Xie

In some implementations, a computer-implemented method of classifying image data includes receiving a plurality of data points corresponding to three-dimensional image data; creating from the plurality of data points a first subset of data points that are above a ground plane in a scene represented by the plurality of data points; identifying a second subset of data points associated with an object in the scene, from the first subset of data points; identifying a plurality of features associated with the second subset of data points and determining a signature for the identified plurality of features; and classifying the second set of data points according to a correspondence between the calculated signature and a reference signature.

Classifying objects in a scene

A method and apparatus for the simplification of a mesh surface is disclosed that preserves the original geometry of the shape of the surface and, at the same time, reduces undesirable triangle geometries. In one embodiment, a mesh simplification process first determines whether an edge swap operation should be performed as a function of a threshold criteria. Such a threshold may be a function of the span angles and cross angles associated with an edge or, alternatively, may be a predetermined span angle size threshold. In another embodiment, the decision as to whether to contract an edge is made by comparing the size of at least one span angle with a span angle threshold and by comparing the sizes of incident angles associated with the edge to an incident angle threshold.

Method and apparatus for fast and efficient mesh simplification

Growing community of inventors

Elmhurst, NY, United States of America

Lingyun Liu