US 11,869,152 B2
Generation of product mesh and product dimensions from user image data using deep learning networks
Chong Jin Koh, Las Vegas, NV (US); Kyohei Kamiyama, Tokyo (JP); and Nobuyuki Hayashi, Tokyo (JP)
Assigned to Bodygram, Inc., New York, NY (US)
Appl. No. 17/924,994
Filed by Bodygram, Inc., New York, NY (US)
PCT Filed May 11, 2021, PCT No. PCT/US2021/070533
§ 371(c)(1), (2) Date Nov. 12, 2022,
PCT Pub. No. WO2021/232049, PCT Pub. Date Nov. 18, 2021.
Claims priority of provisional application 63/024,353, filed on May 13, 2020.
Prior Publication US 2023/0186567 A1, Jun. 15, 2023
Int. Cl. G06T 19/00 (2011.01); G06T 7/60 (2017.01); G06Q 30/0601 (2023.01); G06T 17/20 (2006.01)
CPC G06T 19/00 (2013.01) [G06Q 30/0621 (2013.01); G06Q 30/0643 (2013.01); G06T 7/60 (2013.01); G06T 17/20 (2013.01); G06T 2207/20081 (2013.01); G06T 2207/30196 (2013.01); G06T 2219/012 (2013.01)] 25 Claims
OG exemplary drawing
 
1. A computer-implemented method for generating a 3D mesh model of a product for a user, the computer-implemented method executable by a hardware processor, the method comprising:
receiving one or more images containing a body part of the user;
extracting a body part mesh of the body part from the one or more images, wherein the body part mesh comprises a plurality of body part key points of the body part, and wherein the extracting utilizes a key point deep learning module for generating the body part mesh from the one or more images;
identifying a product mesh subset of the plurality of body part key points from the extracted body part mesh utilizing a product mesh subset deep learning module that has been trained on a product mesh subset training data set,
wherein the product mesh subset training data set comprises one or more sample body part meshes for one or more sample users and a set of sample product mesh subsets for the one or more sample users, and
wherein the product mesh subset deep learning module takes as input the extracted body part mesh; and
generating a product mesh based on the identified product mesh subset of the plurality of body part key points utilizing a product mesh deep learning module that has been trained on a product mesh training data set,
wherein the product mesh training data set comprises one or more sample product mesh subsets for one or more sample users and a set of sample product meshes for the one or more sample users, and
wherein the product mesh deep learning module takes as input the identified product mesh subset.