Patent №
US 6,897,863
Granted
2005-05-24
Filed 2001
Owner
CATERPILLAR INC.
Lab
—
AI components
1
vision
Assignment
Recorded
Dataset
AIPD
2023_r1 edition
Application
09997030
A computer-implemented method is provided for identifying hidden and visible surfaces on an n-dimensional object. An n-dimensional image of an object is generated, the image including a first plurality of n-dimensional components that define a shape and orientation of the image and a plurality of parts located inside the image. An n-dimensional grid of pixels is superimposed on the image. The pixels are arranged in a lattice structure such that each pixel of the grid corresponds to one of a plurality of vertices of an m-sided cell, wherein each side of the m-sided cell includes at least four vertices. A second plurality of n-dimensional components located on an outer perimeter of the image that abut or overlap at least one side of one m-sided cell in the grid is then identified. The second plurality of n-dimensional components constitute the components of the n-dimensional object that are hidden and need not be repeatedly re-rendered.
AI classification
Ownership
CATERPILLAR INC.
assignment · 123450408
Assignors
KINDRATENKO, VOLODYMYR V., FENWICK, ROBERT J.
On an employer assignment, the assignors are typically the inventors.