Nvidia Corporation

Curtis A Beeson

Stephen Donald Lew

Gary D Hicok

Rui Bastos

Michael Brian Cox

John R Zedlewski

Rev Lebaredian

Tony Tamasi

Greg Heinrich

Clement Farabet

Claire Delaunay

Joseph E Demers

Zachary Taylor

Mark Daly

Matthew Campbell

David Auld

In various examples, physical sensor data may be generated by a vehicle in a real-world environment. The physical sensor data may be used to train deep neural networks (DNNs). The DNNs may then be tested in a simulated environment&#x2014;in some examples using hardware configured for installation in a vehicle to execute an autonomous driving software stack&#x2014;to control a virtual vehicle in the simulated environment or to otherwise test, verify, or validate the outputs of the DNNs. Prior to use by the DNNs, virtual sensor data generated by virtual sensors within the simulated environment may be encoded to a format consistent with the format of the physical sensor data generated by the vehicle.

Training, testing, and verifying autonomous machines using simulated environments

A graphical processing unit (GPU) and methods for rendering a three-dimensional (3D) scene generated in a field of view having in-focus and out-of-focus regions on a two-dimensional (2D) screen region of pixels are described. One method includes initially rendering the scene to create color and depth texture maps and creating mip-map layers for the color texture map. The method further comprises subsequently rendering the scene by, for each pixel: creating a mip-map layer selection value as a function of a depth of the pixel from the depth texture map, generating a color value by interpolation using color values from at least one of the mip-map layers chosen according to the mip-map layer selection value, and setting a color of the pixel to the generated color texture.

Depth-of-field effects using texture lookup

Growing community of inventors

Fremont, CA, United States of America

Curtis A Beeson