US 11,720,995 B2
Image rectification
Praveen Narayanan, San Jose, CA (US); Ramchandra Ganesh Karandikar, Palo Alto, CA (US); Nikita Jaipuria, Union City, CA (US); Punarjay Chakravarty, Campbell, CA (US); and Ganesh Kumar, Santa Clara, CA (US)
Assigned to Ford Global Technologies, LLC, Dearborn, MI (US)
Filed by Ford Global Technologies, LLC, Dearborn, MI (US)
Filed on Jun. 4, 2021, as Appl. No. 17/338,900.
Prior Publication US 2022/0392014 A1, Dec. 8, 2022
Int. Cl. G06K 9/00 (2022.01); G06T 3/00 (2006.01); G06T 5/00 (2006.01); G06T 7/00 (2017.01); B60W 10/06 (2006.01); B60W 10/18 (2012.01); B60W 10/20 (2006.01); B60W 10/08 (2006.01); G05B 13/02 (2006.01); B60W 50/06 (2006.01)
CPC G06T 3/0018 (2013.01) [B60W 10/06 (2013.01); B60W 10/08 (2013.01); B60W 10/18 (2013.01); B60W 10/20 (2013.01); B60W 50/06 (2013.01); G05B 13/027 (2013.01); G06T 5/006 (2013.01); G06T 7/0002 (2013.01); B60W 2420/42 (2013.01); G06T 2207/20081 (2013.01); G06T 2207/20084 (2013.01); G06T 2207/30252 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A computer, comprising:
a processor; and
a memory, the memory including instructions executable by the processor to:
input a fisheye image to a vector quantized variational autoencoder;
encode the fisheye image to first latent variables using an encoder;
quantize the first latent variables to generate second latent variables based on an embedded dictionary;
decode the second latent variables to a rectified rectilinear image using a decoder; and
output the rectified rectilinear image.