US 12,229,718 B2
On-shelf image based out-of-stock detection
Francois Chaubard, Burlingame, CA (US)
Assigned to Focal Systems, Inc., Burlingame, CA (US)
Filed by Focal Systems, Inc., Burlingame, CA (US)
Filed on Aug. 2, 2023, as Appl. No. 18/229,574.
Application 18/229,574 is a continuation of application No. 17/173,028, filed on Feb. 10, 2021, granted, now 11,763,254.
Application 17/173,028 is a continuation of application No. 16/024,519, filed on Jun. 29, 2018, granted, now 10,949,799, issued on Mar. 16, 2021.
Prior Publication US 2023/0376896 A1, Nov. 23, 2023
This patent is subject to a terminal disclaimer.
Int. Cl. G06Q 10/087 (2023.01); G06T 11/20 (2006.01); G06V 30/10 (2022.01); G06V 30/14 (2022.01); G06V 30/146 (2022.01); H04N 23/54 (2023.01); H04N 23/698 (2023.01)
CPC G06Q 10/087 (2013.01) [G06T 11/203 (2013.01); G06V 30/1473 (2022.01); H04N 23/54 (2023.01); H04N 23/698 (2023.01); G06T 2210/12 (2013.01); G06V 30/10 (2022.01)] 20 Claims
OG exemplary drawing
 
1. A method for performing a stock detection process comprising:
capturing, using a camera mounted in view of a shelf unit, an image of a product display area of the shelf unit, the image including at least a plurality of a given same product sitting on the shelf unit and a different product from the given same product sitting on the shelf unit, wherein image data of the image includes a bounding box that encompasses the plurality of the given same product;
providing the bounding box as input to a classifier model;
receiving, as output from the classifier model, an indication that the given same product is in a restock state, wherein the classifier model is a machine learning model that is trained to output that the given same product is in a restock state based on a historical rate at which the given same product is removed from the shelf unit; and
responsive to receiving the output, sending a notification that the given same product is determined to be in the restock state.