公开(公告)号：US20220067390A1

公开(公告)日：2022-03-03

申请号：US17463419

申请日：2021-08-31

Applicant: LG ELECTRONICS INC.

Inventor： Amir Hossein KHALILI ,  Bhooshan SUPE ,  Jung Ick GUACK ,  Shantanu PATEL ,  Gaurav SARAF ,  Baisub LEE ,  Helder SILVA ,  Julie HUYNH ,  Jaigak SONG

IPC: G06K9/00 ,  G06T7/246 ,  G06T7/73 ,  G06N3/02

Abstract: Disclosed is a method for identifying and monitoring a shopping behavior in a user. The method includes capturing images from a depth camera mounted on a shelf unit, identifying a user from the captured image, identifying joints of the identified user by performing a deep neural network (DNN) body joint detection on the captured images; detecting and tracking actions of the identified user over a first time period; tracking an object from the bins over a second time period by associating the object with one or more joints among the identified joints that have entered the bins within the shelf unit, and determining an action of the identified user based at least in part on the associated object with the one or more joints and results from the deep learning identification on the bounding box.

公开(公告)号：US20220067688A1

公开(公告)日：2022-03-03

申请号：US17463450

申请日：2021-08-31

Applicant: LG ELECTRONICS INC.

Inventor： Shantanu PATEL ,  Jung Ick Guack ,  Gaurav Saraf ,  Baisub Lee ,  Helder Silva ,  Julie Huynh ,  Jaigak Song ,  Bhooshan Supe ,  Amir Hossein Khalili

IPC: G06Q20/20 ,  G06N20/00 ,  G06Q10/08

Abstract: Disclosed is a method for training a device using machine learning. The method includes obtaining a sensor weight data stream and an environment sensor data stream; determining a new weight value by multiplying a weight measurement among the obtained sensor weight data stream with an environment calibrating coefficient, wherein the environment calibrating coefficient is obtained from using an environment measurement and a pre-recorded calibration file; determining whether a detected change in weight from the device exceeds a predetermined value; and based on a determination that the detected change in weight exceeds the predetermined value, determining a difference in value using the determined new weight value, and identify an item based on inputting the determined difference in value into a machine learning model, wherein the machine learning model is trained to learn unit weights of a particular item to determine an identity of the particular item.