US 11,704,518 B1
Per-image print setting optimization
Sherif Soliman, Seattle, WA (US); Kevin Tsukahara, Frisco, TX (US); Erin Fern Breslin, Seattle, WA (US); Rhia Bucklin, Seattle, WA (US); Val Fox, Shoreline, WA (US); Ron Christopher Belmarch, Seattle, WA (US); and Nick M. Stangel, Frisco, TX (US)
Assigned to AMAZON TECHNOLOGIES, INC., Seattle, WA (US)
Filed by Amazon Technologies, Inc., Seattle, WA (US)
Filed on May 17, 2022, as Appl. No. 17/746,641.
Int. Cl. H04N 1/04 (2006.01); G06K 15/02 (2006.01); G06N 3/08 (2023.01)
CPC G06K 15/1805 (2013.01) [G06K 15/1806 (2013.01); G06N 3/08 (2013.01)] 20 Claims
OG exemplary drawing
 
1. A method, comprising:
receiving first input image data;
generating, using a convolutional neural network, a feature representation of the first input image data;
inputting the feature representation into a classifier network;
generating, by the classifier network using the feature representation, first category data representing a predicted image category for the first input image data;
determining, using a first code associated with first print media, first data representing a color of the first print media and second data representing a fabric type of the first print media;
searching a first data structure using a combination of the first category data, the first data, and the second data as a search query;
determining first printer configuration data stored in the first data structure, wherein the first printer configuration data is associated with the first category data, the color of the first print media, and the fabric type of the first print media by the first data structure, wherein the first printer configuration data comprises a plurality of printer settings for a first printer;
sending the first input image data to a first printer; and
sending the first printer configuration data to the first printer, wherein the first printer is effective to print the first input image data using the first printer configuration data.
 
18. A method comprising:
receiving first input image data;
generating first data representing the first input image data;
generating, by a computer-implemented classifier network based at least in part on the first data, first printer configuration data representing predicted printer settings for the first input image data and for a first printer;
sending the first input image data to a first printer; and
sending the first printer configuration data to the first printer, wherein the first printer is effective to print the first input image data using settings specified by the first printer configuration data.