Electrical tomography-based object recognition

Patent metadata

Field	Value
Publication number	US-11890550-B2
Application number	US-202017132104-A
Country	US
Kind code	B2
Filing date	Dec 23, 2020
Priority date	Jan 2, 2020
Publication date	Feb 6, 2024
Grant date	Feb 6, 2024

Abstract

Improved techniques for identifying objects, such as toy objects, are presented herein. In particular, a play system (e.g., game, doll playset, track set, etc.) includes a probed surface for use in obtaining electrical tomography measurements when toy objects are placed on the probed surface. The play system is configured to analyze the electrical tomography measurements to determine an identity of the toy object placed on the probed surface.

First claim

What is claimed is:

1. A method, comprising:
performing a plurality of electrical tomography measurements for a play surface of a toy; and
analyzing the plurality of electrical tomography measurements to determine an identity of a toy object positioned on the play surface of the toy, wherein the analyzing includes identifying one of:
a stamp on a bottom of the toy object that indicates the identity of the toy object, wherein the stamp is configured to disrupt the plurality of electrical tomography measurements in a known pattern; or
a metallic stamping of a bottom of the toy object that indicates the identity of the toy object.

2. The method of claim 1 , wherein the identity of the toy object is determined utilizing a trained machine learning process.

3. The method of claim 2 , wherein the trained machine learning process is associated with a Convolutional Neural Network.

4. The method of claim 1 , wherein the play surface of the toy is a non-planar play surface of the toy.

5. The method of claim 1 , wherein the play surface of the toy comprises a plurality of electrical probes configured along one or more outer edges of the play surface of the toy.

6. The method of claim 1 , wherein the play surface of the toy comprises a plurality of electrical probes configured on a body of the play surface of the toy.

7. The method of claim 1 , wherein the identity indicates a category associated with the toy object.

8. The method of claim 1 , wherein the identity indicates a model number associated with the toy object.

9. One or more non-transitory computer readable storage media encoded with software comprising computer executable instructions that, when executed by a processor, cause the processor to perform operations, comprising:
performing a plurality of electrical tomography measurements for a play surface of a toy; and
analyzing the plurality of electrical tomography measurements to determine an identity of a toy object positioned on the play surface of the toy, wherein the analyzing includes identifying one of:
a stamp on a bottom of the toy object that indicates the identity of the toy object, wherein the stamp is configured to disrupt the plurality of electrical tomography measurements in a known pattern; or
a metallic stamping of a bottom of the toy object that indicates the identity of the toy object.

10. The media of claim 9 , wherein the identity of the toy object is determined utilizing a trained machine learning process.

11. The media of claim 10 , wherein the trained machine learning process is associated with a Convolutional Neural Network.

12. The media of claim 9 , wherein the identity indicates a category associated with the toy object.

13. The media of claim 9 , wherein the identity indicates a model number associated with the toy object.

14. A toy, comprising:
a play surface;
a memory; and
at least one processor coupled to the play surface and the memory, wherein the processor is configured to:
obtain a plurality of electrical tomography measurements from the play surface, and
analyze the plurality of electrical tomography measurements to determine an identity of a toy object positioned on the play surface, wherein the analyzing includes identifying one of:
a stamp on a bottom of the toy object that indicates the identity of the toy object, wherein the stamp is configured to disrupt the plurality of electrical tomography measurements in a known pattern; or
a metallic stamping of a bottom of the toy object that indicates the identity of the toy object.

15. The toy of claim 14 , wherein the play surface comprises a plurality of electrical probes configured along an outer edge of the play surface.

16. The toy of claim 14 , wherein the play surface comprises a plurality of electrical probes configured on a body of the play surface.

17. The toy of claim 14 , wherein the play surface is a non-planar play surface.

18. The toy of claim 14 , wherein the play surface comprises a plurality of electrical probes configured along one or more outer edges of the play surface.

19. The toy of claim 14 , wherein the identity indicates a category associated with the toy object.

20. The toy of claim 14 , wherein the identity indicates a model number associated with the toy object.

Assignees

• Mattel Inc

Inventors

• Chen Paul

Classifications

• G06N3/0464

  Convolutional networks [CNN, ConvNet] · CPC title

• G06N3/09

  Supervised learning · CPC title

• A63F3/00643 Primary

  Electric board games; Electric features of board games (electric word or number games A63F3/0421; computer chess G06F; electric raffle games A63F3/081) · CPC title

• A63F9/24

  Games using electronic circuits not otherwise provided for · CPC title

• G06F18/00

  Pattern recognition · CPC title