US 10,891,471 B2
Method and system for pose estimation
Xiaogang Wang, Shatin N.T. (CN); Xiao Chu, Shatin N.T. (CN); Wanli Ouyang, Shatin N.T. (CN); and Hongsheng Li, Shatin N.T. (CN)
Assigned to Beijing SenseTime Technology Development Co., Ltd, Haidian District (CN)
Filed by Beijing SenseTime Technology Development Co., Ltd, Haidian District (CN)
Filed on Jan. 2, 2019, as Appl. No. 16/238,347.
Application 16/238,347 is a continuation of application No. 16/089,590, previously published as PCT/CN2016/077505, filed on Mar. 28, 2016.
Prior Publication US 2019/0138799 A1, May 9, 2019
This patent is subject to a terminal disclaimer.
Int. Cl. G06K 9/00 (2006.01); G06K 9/46 (2006.01); G06K 9/62 (2006.01)
CPC G06K 9/00335 (2013.01) [G06K 9/00362 (2013.01); G06K 9/469 (2013.01); G06K 9/6262 (2013.01)] 15 Claims
OG exemplary drawing
 
1. A method for pose estimation, comprising:
extracting a plurality of sets of part-feature maps from an image, wherein each set of the extracted part-feature maps represents a body part and forms a node of a part-feature network;
passing a message of each set of the extracted part-feature maps through the part-feature network to update the extracted part-feature maps, resulting in each set of the extracted part-feature maps incorporating the message of upstream nodes, wherein the passing of the message is performed twice in opposite directions and each pairs of the updated part-feature maps performed in different directions are combined into a score map; and
estimating, based on the combined score maps, the body part within the image.