US 12,494,073 B2
Object labeling in images using dense depth maps
Ilan Tsafrir, Shoham (IL); Guy Tsafrir, Zikhron-Yaakov (IL); Ehud Spiegel, Petach-Tikva (IL); and Dan Atsmon, Rehovot (IL)
Assigned to Cognata Ltd., Rehovot (IL)
Appl. No. 18/022,556
Filed by Cognata Ltd., Rehovot (IL)
PCT Filed Sep. 14, 2021, PCT No. PCT/IL2021/051130
§ 371(c)(1), (2) Date Feb. 22, 2023,
PCT Pub. No. WO2022/054069, PCT Pub. Date Mar. 17, 2022.
Claims priority of provisional application 63/077,729, filed on Sep. 14, 2020.
Prior Publication US 2023/0316789 A1, Oct. 5, 2023
Int. Cl. G06V 20/70 (2022.01); G06V 10/77 (2022.01); G06V 20/58 (2022.01)
CPC G06V 20/70 (2022.01) [G06V 10/7715 (2022.01); G06V 20/58 (2022.01)] 20 Claims
OG exemplary drawing
 
1. A method for annotating a plurality of digital images, comprising:
generating, from a plurality of digital images and a plurality of dense depth maps, each associated with one of the plurality of digital images, an aligned three-dimensional stacked scene representation of a scene, where the plurality of digital images are captured by at least one sensor at the scene, and where each point in the three-dimensional stacked scene is associated with a stability score indicative of a likelihood the point is associated with a static object of the scene;
removing from the three-dimensional stacked scene a plurality of instable points to produce a static three-dimensional stacked scene;
detecting in at least one image of the plurality of digital images at least one static object according to the static three-dimensional stacked scene; and classifying and annotating the at least one static object.