# US 11,442,430 B2 · Rapid packaging prototyping using machine learning

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 11,442,430 B2
- Document ID: US11442430B2
- Status: Granted
- Kind: B2
- Application number: 16/552,946
- Filed: 2019-08-27
- Published: 2021-03-04
- Granted: 2022-09-13
- Earliest claimed: 2019-08-27
- Assignees: KYOCERA Document Solutions Inc.
- Applicants: KYOCERA Document Solutions, Inc.
- Inventors: Eric de Beus
- Web page: https://patentgrep.com/patent/US11442430B2

## CPC

- G05B 19/4099: Surface or curve machining, making three-dimensional [3D] objects, e.g. desktop manufacturing
- G06N 3/045: Combinations of networks
- G06N 3/0464: Convolutional networks [CNN, ConvNet]
- G06N 3/08: Learning methods
- G06N 3/084: Backpropagation, e.g. using gradient descent
- G06N 3/09: Supervised learning
- G06N 20/10: using kernel methods, e.g. support vector machines [SVM]
- G05B 2219/49017: DTM desktop manufacturing, prototyping
- G05B 2219/49023: 3-D printing, layer of powder, add drops of binder in layer, new powder
- G06N 7/01: Probabilistic graphical models, e.g. probabilistic networks
- G05B: CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS (systems for controlling or regulating non-electric variables G05D; systems for regulating electric or magnetic variables G05F; control devices or systems insofar as characterised by mechanical features only G05G)
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS

## Abstract

A system includes a scanner to convert a packaging sketch into a pixelated image and a convolutional neural network configured to segment the pixelated image into bounded objects including fold lines and cut lines. A controller is configured to transform the fold lines and the cut lines into control commands to a folding machine and a cutting machine. A method includes converting a packaging sketch into a pixelated image using a scanner and segmenting, using a convolutional neural network, the pixelated image into bounded objects including fold lines and cut lines. The method also includes transforming, using a controller, the fold lines and the cut lines into control commands to a folding machine and a cutting machine.

## Claims

### Claim 1

A system comprising:
a scanner to convert a packaging sketch into a pixelated image;
a convolutional neural network configured to segment the pixelated image into bounded objects including fold lines and cut lines; and
a controller configured to transform the fold lines and the cut lines into control commands to a folding machine and a cutting machine.

### Claim 2 (depends on 1)

The system of claim 1, wherein the convolutional neural network is further configured to generate package design instructions based on the pixelated image that has been segmented.

### Claim 3 (depends on 2)

The system of claim 2, further comprising an operator machine interface, wherein before the controller receives the package design instructions, the package design instructions are rendered as a displayable package design drawing and displayed on the operator machine interface.

### Claim 4 (depends on 3)

The system of claim 3, further comprising logic to:
on condition the displayable package design drawing does not require changes, send the package design instructions to the controller; and
on condition the displayable package design drawing requires changes, convert an altered package design drawing into an updated pixelated image using the scanner.

### Claim 5 (depends on 1)

The system of claim 1, further comprising:
logic to apply annotations associated with the bounded objects to transform non-similar bounded objects into similar bounded objects.

### Claim 6 (depends on 5)

The system of claim 5, further comprising:
logic to alter the fold lines and the cut lines to transform the non-similar bounded objects into the similar bounded objects.

### Claim 7 (depends on 1)

The system of claim 1, wherein the fold lines are depicted by solid lines between the bounded objects on the packaging sketch.

### Claim 8 (depends on 1)

The system of claim 1, wherein at least some of the cut lines are perforation lines depicted by irregular lines on the packaging sketch.

### Claim 9 (depends on 1)

The system of claim 1, wherein the cut lines are computed from a machine-derived rectangular bounding box for the packaging sketch.

### Claim 10 (depends on 1)

The system of claim 1, further comprising a feedback path from a printer to the scanner.

### Claim 11

A method comprising:
converting a packaging sketch into a pixelated image using a scanner;
segmenting, using a convolutional neural network, the pixelated image into bounded objects including fold lines and cut lines; and
transforming, using a controller, the fold lines and the cut lines into control commands to a folding machine and a cutting machine.

### Claim 12 (depends on 11)

The method of claim 11, further comprising generating package design instructions based on the pixelated image that has been segmented.

### Claim 13 (depends on 12)

The method of claim 12, further comprising, before the transforming step, rendering the package design instructions as a displayable package design drawing and displaying the displayable package design drawing on an operator machine interface.

### Claim 14 (depends on 13)

The method of claim 13, further comprising:
on condition the displayable package design drawing does not require changes, initiating the transforming step; and
on condition the displayable package design drawing requires changes, altering the displayable package design drawing and reinitiating at least one of the converting, segmenting, and generating steps with the altered drawing.

### Claim 15 (depends on 11)

The method of claim 11, further comprising using logic to apply annotations associated with the bounded objects to transform non-similar bounded objects into similar bounded objects.

### Claim 16 (depends on 15)

The method of claim 15, further comprising using logic to alter the fold lines and the cut lines to transform the non-similar bounded objects into the similar bounded objects.

### Claim 17 (depends on 11)

The method of claim 11, wherein the fold lines are depicted by solid lines between the bounded objects on the packaging sketch.

### Claim 18 (depends on 11)

The method of claim 11, wherein at least some of the cut lines are perforation lines depicted by irregular lines on the packaging sketch.

### Claim 19 (depends on 11)

The method of claim 11, wherein the cut lines are computed from a machine-derived rectangular bounding box for the packaging sketch.

### Claim 20 (depends on 11)

The method of claim 11, further comprising providing feedback through a feedback path from a printer to the scanner.