# US 2023/0267684 A1 · Generating training data through image augmentation

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

- Number: US 2023/0267684 A1
- Document ID: US20230267684A1
- Status: Published application
- Kind: A1
- Application number: 17/676,444
- Filed: 2022-02-21
- Published: 2023-08-24
- Earliest claimed: 2022-02-21
- Assignees: International Business Machines Corporation
- Applicants: International Business Machines Corporation
- Inventors: Kun Yan Yin; Zhong Fang Yuan; Yi Chen Zhong; Lu Yu; Tong Liu
- Web page: https://patentgrep.com/patent/US20230267684A1

## CPC

- G06T 17/10: Constructive solid geometry [CSG] using solid primitives, e.g. cylinders, cubes
- G06T 19/00: Manipulating three-dimensional [3D] models or images for computer graphics
- G06F 16/532: Query formulation, e.g. graphical querying
- G06N 3/0455: Auto-encoder networks; Encoder-decoder networks
- G06N 3/0475: Generative networks
- G06N 3/08: Learning methods
- G06N 3/094: Adversarial learning
- G06V 10/764: using classification, e.g. of video objects
- G06V 10/82: using neural networks
- G06V 20/64: Three-dimensional [3D] objects
- G06N 10/60: Quantum algorithms, e.g. based on quantum optimisation, quantum Fourier or Hadamard transforms
- G06T: IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06V: IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING

## Abstract

An example operation may include one or more of generating a three-dimensional (3D) model of an object via execution of a machine learning model on one or more images of the object, capturing a plurality of snapshots of the 3D model of the object at different angles to generate a plurality of snapshot images of the object, fusing a feature into each of the plurality of snapshots to generate a plurality of fused snapshots of the 3D model of the object, and storing the plurality of fused snapshots of the 3D model of the object in memory.

## Claims

### Claim 1

An apparatus comprising:
a processor configured to
generate a three-dimensional (3D) model of an object via execution of a machine learning model on one or more images of the object,
capture a plurality of snapshots of the 3D model of the object at different angles to generate a plurality of snapshot images of the object,
fuse a feature into each of the plurality of snapshots to generate a plurality of fused snapshots of the 3D model of the object, and
store the plurality of fused snapshots of the 3D model of the object in memory.

### Claim 2 (depends on 1)

The apparatus of claim 1, wherein the processor is further configured to segment the one or more images of the object to generate one or more segmented images in which an area around the object is removed, and generate the 3D model of the object based on the one or more segmented images.

### Claim 3 (depends on 1)

The apparatus of claim 1, wherein the processor is configured to execute a machine learning model and input the plurality of fused snapshots into the executing machine learning model to train the machine learning model.

### Claim 4 (depends on 1)

The apparatus of claim 1, wherein the processor is configured to generate the 3D model of the object via execution of a plurality of autoencoders which extract different attributes of the object from the one or more images.

### Claim 5 (depends on 4)

The apparatus of claim 4, wherein the processor is configured to input the outputs of the plurality of autoencoders into a Style Generative Adversarial Network (StyleGAN) to generate the 3D model of the object.

### Claim 6 (depends on 1)

The apparatus of claim 1, wherein the processor is further configured to receive a string input entered via a user interface and retrieve a plurality of images of the object from an image-based search engine via the string input.

### Claim 7 (depends on 6)

The apparatus of claim 6, wherein the processor is configured to identify a subject, a predicate, and an object from the string, and query the image-based search engine based on a relationship between the subject, the predicate, and the object.

### Claim 8 (depends on 1)

The apparatus of claim 1, wherein the processor is configured to fuse the feature into the plurality of snapshot images via execution of an Auto Generative Adversarial Network (AutoGAN).

### Claim 9

A method comprising:
generating a three-dimensional (3D) model of an object via execution of a machine learning model on one or more images of the object;
capturing a plurality of snapshots of the 3D model of the object at different angles to generate a plurality of snapshot images of the object;
fusing a feature into each of the plurality of snapshots to generate a plurality of fused snapshots of the 3D model of the object; and
storing the plurality of fused snapshots of the 3D model of the object in memory.

### Claim 10 (depends on 9)

The method of claim 9, wherein the method further comprises segmenting the one or more images of the object to generate one or more segmented images in which an area around the object is removed, and generating the 3D model of the object based on the one or more segmented images.

### Claim 11 (depends on 9)

The method of claim 9, wherein the method further comprises executing a machine learning model and inputting the plurality of fused snapshots into the executing machine learning model to train the machine learning model.

### Claim 12 (depends on 9)

The method of claim 9, wherein the generating comprises generating the 3D model of the object via execution of a plurality of autoencoders which extract different attributes of the object from the one or more images.

### Claim 13 (depends on 12)

The method of claim 12, wherein the generating comprises inputting the outputs of the plurality of autoencoders into a Style Generative Adversarial Network (StyleGAN) to generate the 3D model of the object.

### Claim 14 (depends on 9)

The method of claim 9, wherein the method further comprises receiving a string input entered via a user interface and retrieve a plurality of images of the object from an image-based search engine via the string input.

### Claim 15 (depends on 14)

The method of claim 14, wherein the method further comprises identifying a subject, a predicate, and an object from the string, and querying the image-based search engine based on a relationship between the subject, the predicate, and the object.

### Claim 16 (depends on 9)

The method of claim 9, wherein the fusing comprises fusing the feature into the plurality of snapshot images via execution of an Auto Generative Adversarial Network (AutoGAN).

### Claim 17

A computer-readable storage medium comprising instructions, that when read by a processor, cause the processor to perform a method comprising:
generating a three-dimensional (3D) model of an object via execution of a machine learning model on one or more images of the object;
capturing a plurality of snapshots of the 3D model of the object at different angles to generate a plurality of snapshot images of the object;
fusing a feature into each of the plurality of snapshots to generate a plurality of fused snapshots of the 3D model of the object; and
storing the plurality of fused snapshots of the 3D model of the object in memory.

### Claim 18 (depends on 17)

The computer-readable storage medium of claim 17, wherein the method further comprises segmenting the one or more images of the object to generate one or more segmented images in which an area around the object is removed, and generating the 3D model of the object based on the one or more segmented images.

### Claim 19 (depends on 17)

The computer-readable storage medium of claim 17, wherein the method further comprises executing a machine learning model and inputting the plurality of fused snapshots into the executing machine learning model to train the machine learning model.

### Claim 20 (depends on 17)

The computer-readable storage medium of claim 17, wherein the generating comprises generating the 3D model of the object via execution of a plurality of autoencoders which extract different attributes of the object from the one or more images.
