Home appliance and method for controlling home appliance

Patent metadata

Field	Value
Publication number	US-11634849-B2
Application number	US-202016803597-A
Country	US
Kind code	B2
Filing date	Feb 27, 2020
Priority date	Nov 15, 2019
Publication date	Apr 25, 2023
Grant date	Apr 25, 2023

Abstract

A home appliance and a control method for the home appliance, which is operable in an IoT environment through a 5G communication network and uses a neural network model generated according to machine learning is provided. The home appliance may include a home appliance main body; a container mounted within the home appliance main body to accommodate a treatment target; a camera arranged to photograph the inside of the container; and one or more processors configured to control an operation of the home appliance, wherein the processor is configured to determine an amount of a treatment target based on feature shapes of the container identified in an image of the inside of the container photographed by the camera.

First claim

What is claimed is:

1. A home appliance, comprising:
a main body;
a container mounted within the main body to accommodate a treatment target in an interior of the container;
a camera arranged to photograph the interior of the container;
a memory including a stored neural network model trained to determine an amount of a treatment target by analyzing an image of the inside of the container acquired through the camera; and
one or more processors programmed to control an operation of the home appliance,
wherein a first processor of the one or more processors is programmed to perform an operation to determine an amount of the treatment target based on feature shapes of the container identified in an image of the interior of the container photographed by the camera through the neural network model,
wherein the amount of the treatment target is a volume of the treatment target, and
wherein the feature shapes of the container are based on holes in a bottom surface or a side surface of the container.

2. The home appliance of claim 1 , wherein the first processor determines the amount of the treatment target based on a number of shapes of a first form included in the feature shapes of the container identified in the image of the interior of the container through the pre-trained neural network model.

3. The home appliance of claim 1 , further comprising:
a lighting disposed to illuminate the interior of the container; and
a door configured to open and close a treatment target inlet of the container,
wherein the camera is disposed in the door.

4. The home appliance of claim 1 , wherein the operation of determining the amount of the treatment target based on the feature shapes inside the container comprises operations of
extracting the feature shapes from the image of the interior of the container before the treatment target is put into the container;
correlating the amount of the treatment target with blocked feature shapes or visible feature shapes; and
determining the amount of the treatment target based on the blocked feature shapes or the visible feature shapes in the image of the interior of the container after the treatment target is put into the container.

5. The home appliance of claim 1 , wherein the neural network model is pre-trained using training data comprising images of the interior of the container into which various amounts of the treatment target is put into the container and labels indicating the amount of the treatment target for each image.

6. The home appliance of claim 5 , wherein the neural network model is configured to determine the amount of the treatment target using a number of blocked feature shapes or a number of visible feature shapes among the feature shapes in the container before the treatment target is put into the container.

7. The home appliance of claim 1 , further comprising a weight sensor configured to detect a weight of the treatment target in the container,
wherein the first processor is further configured to determine a density of the treatment target based on the volume of the treatment target and the weight of the treatment target detected by the weight sensor.

8. The home appliance of claim 7 , wherein the first processor is further configured to:
determine a type of the treatment target based on object recognition of an object recognition model for the image of the treatment target photographed from the camera and the density of the treatment target; and
select a treatment mode based on the type of the treatment target.

9. The home appliance of claim 1 , wherein the first processor is further configured to determine at least one of a water supply amount or a detergent input amount based on the amount of the treatment target.

Assignees

• Lg Electronics Inc

Inventors

• Park Yun Sik
• Han dayun
• Hwang Sung Mok

Classifications

• G06N3/09

  Supervised learning · CPC title

• G06N3/0464

  Convolutional networks [CNN, ConvNet] · CPC title

• D06F2105/02

  Water supply · CPC title

• G06N3/092

  Reinforcement learning · CPC title

• G06N3/088

  Non-supervised learning, e.g. competitive learning · CPC title