US 12,254,638 B2
Method and system of recognizing object edges and computer-readable storage medium
Qingsong Xu, Hangzhou (CN); and Qing Li, Hangzhou (CN)
Assigned to Hangzhou Glority Software Limited, Hangzhou (CN)
Appl. No. 17/637,801
Filed by Hangzhou Glority Software Limited, Hangzhou (CN)
PCT Filed Aug. 30, 2021, PCT No. PCT/CN2021/115482
§ 371(c)(1), (2) Date Feb. 24, 2022,
PCT Pub. No. WO2022/057607, PCT Pub. Date Mar. 24, 2022.
Claims priority of application No. 202010995134.7 (CN), filed on Sep. 21, 2020.
Prior Publication US 2023/0267619 A1, Aug. 24, 2023
This patent is subject to a terminal disclaimer.
Int. Cl. G06T 7/13 (2017.01); G06V 10/44 (2022.01); G06V 10/74 (2022.01)
CPC G06T 7/13 (2017.01) [G06V 10/44 (2022.01); G06V 10/761 (2022.01); G06T 2207/20084 (2013.01); G06T 2207/20164 (2013.01)] 14 Claims
OG exemplary drawing
 
1. A method of recognizing object edges, comprising:
obtaining an input image, wherein the input image comprises an object with edges, and the edges of the object comprise a plurality of object vertices, wherein the plurality of object vertices comprise a first object vertex and a second object vertex;
recognizing, through an object vertex recognition model, the input image and obtaining a first relative position between the first object vertex and a first image vertex, and obtaining a second relative position between the second object vertex and a second image vertex;
determining a reference position of each of the object vertices in the input image according to the relative position of each of the object vertices and the corresponding image vertex thereof;
performing a corner point detection in a predetermined area where the reference position of the object vertex is located for each of the object vertices;
determining an actual position of each of the object vertices in the input image according to a result of the corner point detection; and
sequentially connecting adjacent object vertices to form edge lines to obtain the edges of the object with edges in the input image according to the actual position of each of the object vertices in the input image.