US 12,272,163 B2
Unsupervised object-oriented decompositional normalizing flow
Farhad Ghazvinian Zanjani, Almere (NL); Hanno Ackermann, Hilversum (NL); Daniel Hendricus Franciscus Dijkman, Haarlem (NL); and Fatih Murat Porikli, San Diego, CA (US)
Assigned to QUALCOMM Incorporated, San Diego, CA (US)
Filed by QUALCOMM Incorporated, San Diego, CA (US)
Filed on Jul. 7, 2022, as Appl. No. 17/859,913.
Claims priority of provisional application 63/303,789, filed on Jan. 27, 2022.
Prior Publication US 2023/0237819 A1, Jul. 27, 2023
Int. Cl. G06V 10/82 (2022.01); G06T 7/194 (2017.01); G06V 20/70 (2022.01)
CPC G06V 20/70 (2022.01) [G06T 7/194 (2017.01); G06V 10/82 (2022.01); G06T 2207/20081 (2013.01); G06T 2207/20084 (2013.01)] 30 Claims
OG exemplary drawing
 
19. A processor-implemented method for processing data, the method comprising:
determining, based on processing data depicting multiple targets in a scene using an scene-decompositional model having a normalizing flow neural network architecture, a distribution of scene data as a mixture of flows from one or more background components and one or more foreground components;
processing the distribution of scene data using the scene-decompositional model;
identifying, based on the processing of the distribution of the scene data using the scene-decompositional model, a target associated with the one or more foreground components and included in the data depicting the multiple targets in the scene; and
outputting a representation of the target.