Granted Patent B2

US 12,555,435 · App. 18/397,525 · Granted Feb 17, 2026

Inventors: Bryan Kelly (Rancho Santa Margarita, CA); Martin S. Lyons (Henderson, NV)

Assignee: LNW Gaming, Inc.

G07F17/3227G06V10/764G06V10/82G06V40/10G06V40/172G06V40/20G07F17/3206H04N7/18H04N23/62H04N23/80

Abstract

Claims (55)

1 . A method of operating a wagering game system comprising a gaming table and a camera, said method comprising:

detecting, by an electronic processor in response to electronic communication with the wagering game system, a game state from a plurality of game states associated with a wagering game presented at the gaming table;

dynamically generating, by the electronic processor based at least in part on a graphical template associated with the game state, a set of digital images cropped, using the template, from portions of a frame of image data captured by the camera at the gaming table;

determining, by the electronic processor via electronic analysis of the set of digital images, that a collective size of the set of digital images as a unit meets a maximum resolution target input requirement for a machine learning model associated with the game state; and

in response to determining that the collective size of the set of digital images as a unit meets the maximum resolution target input requirement, providing, by the electronic processor, the set of digital images as a single file to the machine learning model for concurrent analysis of the set of digital images.

2 . The method of claim 1 , wherein in response to the detecting the game state, selecting, from a plurality of templates associated with the plurality of game states, the graphical template.

3 . The method of claim 1 , wherein the dynamically generating the set of digital images comprises:

superimposing, by the electronic processor, the graphical template over an image feed of the gaming table, wherein the graphical template specifies labeled areas associated with areas of interest at the gaming table;

cropping, by the electronic processor using known coordinates for the labeled areas, the set of digital images from the image feed, wherein each of the set of digital images has a resolution of the image feed; and

combining, by the electronic processor, the set of digital images onto a sprite sheet.

4 . The method of claim 1 , wherein prior to determining that the collective size of the set of digital images meets the maximum resolution target input requirement, said method further comprising:

determining that the set of digital images fails the maximum resolution target input requirement for the machine learning model; and

in response to determining that the set of digital images fails the maximum resolution target input requirement, iteratively modifying, by the electronic processor, an image resolution property of at least one image that is largest in size from the set of digital images until the iteratively modifying causes the set of digital images to collectively meet the maximum resolution target input requirement, wherein the iteratively modifying comprises, in an iterative manner until the set of digital images meets the maximum resolution target input requirement, scaling down an image resolution dimension of the at least one image that is largest in size by a pixel-dimension reduction value.

5 . The method of claim 4 , wherein the scaling down the image resolution dimension of the at least one image that is largest in size by the pixel-dimension reduction value comprises, for each iteration of the iteratively modifying, scaling down a larger one of either an image resolution width or an image resolution height of the at least one image that is largest in size by the pixel-dimension reduction value and scaling down a smaller one of either the image resolution width or the image resolution height by the pixel-dimension reduction value divided by an aspect ratio at which the frame of image data was captured.

6 . The method of claim 5 further comprising, running a rectangle packing algorithm on the set of digital images each time the image resolution dimension is scaled down to determine whether the set of digital images collectively fit into a rectangle that represents a maximum resolution limit for the machine learning model.

7 . The method of claim 4 , wherein the pixel-dimension reduction value is one pixel.

8 . The method of claim 4 further comprising:

prior to the iteratively modifying, determining that an amount to which the set of digital images needs to be modified is above a given level;

selecting, based on the determined amount being above the given level, a first value for the pixel-dimension reduction value, wherein the first value is more than one pixel in size;

after occurrence of one or more instances of iteratively modifying the image resolution property using the first value, determining that the amount to which the set of digital images needs to be modified is below the given level; and

in response to determining that the amount to which the set of digital images needs to be modified is below the given level, dynamically decreasing the pixel-dimension reduction value to a second value less than the first value.

9 . A system comprising:

a camera configured to capture an image feed of a gaming table; and

an electronic processor configured to execute instructions, which when executed cause the system to perform operations to

detect a game state of a wagering game presented at the gaming table, wherein the game state is one of a plurality of game states possible for the wagering game;

select, from a plurality of templates associated with the plurality of game states, a graphical template that corresponds to the game state;

superimpose the graphical template over the image feed of the gaming table, wherein the graphical template specifies labeled areas associated with areas of interest at the gaming table;

generate, using known coordinates for the labeled areas, a set of digital images cropped from the image feed, wherein each of the set of digital images has a resolution of the image feed;

determine, in response to iterative analysis of the set of digital images, that a collective size of the set of digital images as a unit meets a maximum resolution target input requirement for a machine learning model associated with the game state;

in response to determination that the collective size of the set of digital images as a unit meets the maximum resolution target input requirement, combine the set of digital images into a sprite sheet; and

provide the sprite sheet to the machine learning model for concurrent analysis of the set of digital images.

10 . The system of claim 9 , wherein the electronic processor is further configured to execute instructions that, when executed, cause the system to perform operations to:

determine that the collective size of the set of digital images as a unit fails the maximum resolution target input requirement;

in response to determination that the set of digital images collectively fails the maximum resolution target input requirement, identify, via analysis of image sizes of the set of digital images, at least one image from the set of digital images that possesses an image resolution dimension that is largest in size amongst all members of the set of digital images; and

iteratively modify the image resolution dimension of the at least one image until the set of digital images collectively meet the maximum resolution target input requirement.

11 . The system of claim 10 , wherein the electronic processor configured to execute instructions to cause the system to perform the operation to iteratively modify the image resolution dimension of the at least one image until the set of digital images collectively meet the maximum resolution target input requirement is further configured to execute instructions that, when executed, cause the system to perform operations to scale down, for each instance of iteratively modifying the image resolution dimension, a resolution property of the image resolution dimension by one pixel until the set of digital images collectively meet the maximum resolution target input requirement.

12 . The system of claim 11 , wherein the electronic processor configured to execute instructions to cause the system to perform the operation to scale down, for each instance of iteratively modifying the image resolution dimension, the resolution property of the image resolution dimension by one pixel until the set of digital images collectively meet the maximum resolution target input requirement is further configured to execute instructions that, when executed, cause the system to perform operations to, for each instance of iteratively modifying the image resolution dimension, scale an image resolution width of the at least one image by the one pixel and scale an image resolution height of the at least one image by one pixel divided by an aspect ratio at which the image feed was captured.

13 . The system of claim 11 , wherein the electronic processor configured to execute instructions to cause the system to perform the operation to scale down, for each iterative modification, the resolution property of the image resolution dimension by one pixel until the set of digital images collectively meet the maximum resolution target input requirement is further configured to execute instructions that, when executed, cause the system to perform operations to, for each instance of iteratively modifying the image resolution dimension, scale a first image resolution dimension of the at least one image by the one pixel and scale a second image resolution dimension of the at least one image by one pixel divided by an aspect ratio at which the image feed was captured.

14 . The system of claim 11 , wherein the electronic processor configured to execute instructions to cause the system to perform the operation to scale down, for each iterative modification, the resolution property of the image resolution dimension by one pixel until the set of digital images collectively meet the maximum resolution target input requirement is further configured to execute instructions that, when executed, cause the system to perform operations to, for each instance of iteratively modifying the image resolution dimension, determine, via a rectangle packing algorithm, whether the set of digital images collectively fit into a rectangle that represents the maximum resolution target input requirement.

15 . The system of claim 9 , wherein a location of each given one of the set of digital images in the sprite sheet is associated with a unique identifier for a respective one of the labeled areas from which each given one of the set of digital images was cropped, and wherein each of set of digital images is positioned on the sprite sheet using each respective unique identifier.

16 . One or more non-transitory, machine-readable media having instructions stored thereon, which instructions, when executed by one or more electronic processors of a wagering game system, cause the wagering game system to perform operations comprising:

detect a game state of a wagering game presented at a gaming table;

dynamically generating, by the electronic processor based at least in part on a graphical template associated with the game state, a set of digital images cropped, using the template, from portions of a frame of image data captured by a camera at the gaming table;

in response to determining, by the electronic processor via electronic analysis of the set of digital images, that the set of digital images fails a maximum resolution target input requirement for a machine learning model that corresponds to the game state, iteratively modifying, by the electronic processor, an image resolution property of at least one image that is largest in size from the set of digital images until the iteratively modifying causes the set of digital images to collectively meet the maximum resolution target input requirement, wherein the iteratively modifying comprises, in an iterative manner, until the set of digital images meets the maximum resolution target input requirement, scaling down an image resolution dimension of the at least one image that is largest in size by a pixel-dimension reduction value; and

in response to determining that said iteratively modifying causes the set of digital images to meet the maximum resolution target input requirement, providing, by the electronic processor, the set of digital images as a single file to the machine learning model for concurrent analysis of the set of digital images.

17 . The one or more non-transitory machine-readable media of claim 16 , wherein the instructions, when executed by the one or more electronic processors, cause the wagering game system to perform operations further comprising:

dynamically decreasing, during the iteratively modifying, the pixel-dimension reduction value in response to determination that an overall resolution size of the set of digital images approaches a required collective resolution size.

18 . The one or more non-transitory machine-readable media of claim 16 , wherein the instructions, when executed by the one or more electronic processors, cause the wagering game system to perform operations further comprising:

dynamically decreasing, during the iteratively modifying, the pixel-dimension reduction value based on an amount of time that the machine learning model needs to complete the concurrent analysis of the set of digital images for the game state.

19 . The one or more non-transitory machine-readable media of claim 16 , wherein the instructions, when executed by the one or more electronic processors, cause the wagering game system to perform operations further comprising:

prior to the iteratively modifying, determining that an amount to which the set of digital images needs to be modified is above a given level;

selecting, based on the determined amount being above the given level, a first value for the pixel-dimension reduction value, wherein the first value is more than one pixel in size;

after occurrence of one or more instances of iteratively modifying the image resolution property using the first pixel-dimension reduction value, determining that the amount to which the set of digital images needs to be modified is below the given level; and

in response to determining that the amount to which the set of digital images needs to be modified is below the given level, dynamically decreasing the pixel-dimension reduction value to a second value less than the first value.

20 . The one or more non-transitory machine-readable media of claim 19 , wherein the first value is at least ten pixels in size, and wherein the second value is one pixel in size.

Assignments (4)

Continuity (3)

References Cited (141)