# US 2022/0284484 A1 · Using machine learning techniques to calculate damage of vehicles involved in an accident

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

- Number: US 2022/0284484 A1
- Document ID: US20220284484A1
- Status: Published application
- Kind: A1
- Application number: 17/826,250
- Filed: 2022-05-27
- Published: 2022-09-08
- Earliest claimed: 2017-08-11
- Assignees: State Farm Mutual Automotive Insurance Company
- Applicants: State Farm Mutual Automotive Insurance Company
- Inventors: Brian Mark Fields; Lee Marvin John Assam
- Web page: https://patentgrep.com/patent/US20220284484A1

## CPC

- G06Q 30/0278: {Product appraisal}
- G06Q 40/08: Insurance
- G06Q: INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR

## Abstract

Disclosed systems and methods incorporate machine learning to assess damage to vehicles. An example method includes: accessing image data representing one or more digital images of damage to a vehicle; selecting, using one or more processors, an applicable machine learning algorithm from a plurality of different trained machine learning algorithms, wherein the plurality of different machine learning algorithms are trained for respective different combinations of one or more of vehicle make, vehicle model, vehicle year, or area of damage; processing the image data, with the applicable machine learning algorithm using one or more processors, to determine assessed damage for the vehicle; accessing actual damage information for the vehicle; determining, using one or more processors, differences between the actual damage and the assessed damage; and iterating, using one or more processors, the applicable machine learning algorithm based on the differences to improve its damage assessment accuracy.

## Claims

### Claim 1

A method, comprising:
accessing image data representing one or more digital images of damage to a vehicle;
selecting, using one or more processors, an applicable machine learning algorithm from a plurality of different trained machine learning algorithms, wherein the plurality of different machine learning algorithms are trained for respective different combinations of one or more of vehicle make, vehicle model, vehicle year, or area of damage;
processing the image data, with the applicable machine learning algorithm using one or more processors, to determine assessed damage for the vehicle;
accessing actual damage information for the vehicle;
determining, using one or more processors, differences between the actual damage and the assessed damage; and
iterating, using one or more processors, the applicable machine learning algorithm based on the differences to improve its damage assessment accuracy.

### Claim 2 (depends on 1)

The method of claim 1, further comprising:
comparing, via one or more processors, the assessed damage to a threshold damage; and
sending, via a computer network, an alert to an emergency service when the assessed damage is above the threshold damage.

### Claim 3 (depends on 1)

The method of claim 1, wherein the one or more images are received from an unmanned aerial vehicle having an imaging sensor.

### Claim 4 (depends on 1)

The method of claim 1, further comprising:
accessing second image data representing one or more digital images of one or more other vehicles with known actual damage; and
processing the image data and the second image data, with the applicable machine learning algorithm using one or more processors, to determine the assessed damage.

### Claim 5 (depends on 4)

The method of claim 4, wherein the applicable machine learning algorithm is trained to process the image data and the second image data by:
determining, using one or more processors, differences between the image data and the second image data; and
processing the differences with the applicable machine learning algorithm.

### Claim 6 (depends on 4)

The method of claim 4, further comprising:
retrieving insurance claim data for the one or more other vehicles; and
processing the image data, the second image data and the insurance claim data, with the applicable machine learning algorithm using one or more processors, to determine the assessed damage.

### Claim 7 (depends on 6)

The method of claim 6, wherein the applicable machine learning algorithm is trained to process the image data, the second image data and the insurance claim data by:
determining, based on the image data and the second image data, one or more components of the vehicle that are damaged; and
determining, based on the one or more components and insurance claim data related to the one or more components, the assessed damage.

### Claim 8 (depends on 1)

The method of claim 1, wherein the assessed damage is indicative of a cost of repair for the vehicle.

### Claim 9 (depends on 1)

The method of claim 1, wherein iterating the applicable machine learning algorithm comprises at least one of:
adjusting, using one or more processors, one or more equations used to assess damage;
adjusting, using one or more processors, one or more variables of the one or more equations used to assess damage; or
adjusting, using one or more processors, one or more weights of the one or more variables of the one or more equations used to assess damage.

### Claim 10

A system, comprising:
one or more processors; and
one or more non-transitory storage media storing instructions that, when executed by the one or more processors, cause the system to:
access image data representing one or more digital images of damage to a vehicle;
select an applicable machine learning algorithm from a plurality of different trained machine learning algorithms, wherein the plurality of different machine learning algorithms are trained for respective different combinations of one or more of vehicle make, vehicle model, vehicle year, or area of damage;
process the image data, with the applicable machine learning algorithm using one or more processors, to determine assessed damage for the vehicle;
access actual damage information for the vehicle;
determine differences between the actual damage and the assessed damage; and
iterate the applicable machine learning algorithm based on the differences to improve its damage assessment accuracy.

### Claim 11 (depends on 10)

The system of claim 10, further comprising a communication interface, and wherein the instructions, when executed by the one or more processors, cause the system to:
compare the assessed damage to a threshold damage; and
send, via the communication interface, an alert to an emergency service when the assessed damage is above the threshold damage.

### Claim 12 (depends on 10)

The system of claim 10, wherein the instructions, when executed by the one or more processors, cause the system to:
access second image data representing one or more digital images of one or more other vehicles with known actual damage; and
process the image data and the second image data, with the applicable machine learning algorithm using one or more processors, to determine the assessed damage.

### Claim 13 (depends on 12)

The system of claim 12, wherein the applicable machine learning algorithm is trained to process the image data and the second image data by:
determining differences between the image data and the second image data; and
processing the differences with the applicable machine learning algorithm.

### Claim 14 (depends on 12)

The system of claim 12, wherein the instructions, when executed by the one or more processors, cause the system to:
retrieve insurance claim data for the one or more other vehicles; and
process the image data, the second image data and the insurance claim data, with the applicable machine learning algorithm using one or more processors, to determine the assessed damage.

### Claim 15 (depends on 14)

The system of claim 14, wherein the applicable machine learning algorithm is trained to process the image data, the second image data and the insurance claim data by:
determining, based on the image data and the second image data, one or more components of the vehicle that are damaged; and
determining, based on the one or more components and insurance claim data related to the one or more components, the assessed damage.

### Claim 16 (depends on 10)

The system of claim 10, wherein the assessed damage is indicative of a cost of repair for the vehicle.

### Claim 17 (depends on 10)

The system of claim 10, wherein the instructions, when executed by the one or more processors, cause the system to iterate the applicable machine learning algorithm by at least one of:
adjusting one or more equations used to assess damage;
adjusting one or more variables of the one or more equations used to assess damage; or
adjusting one or more weights of the one or more variables of the one or more equations used to assess damage.

### Claim 18

A non-transitory computer-readable storage media storing computer-readable instructions that, when executed by one or more processors, cause the one or more processors to:
access image data representing one or more digital images of damage to a vehicle;
select an applicable machine learning algorithm from a plurality of different trained machine learning algorithms, wherein the plurality of different machine learning algorithms are trained for respective different combinations of one or more of vehicle make, vehicle model, vehicle year, or area of damage;
process the image data, with the applicable machine learning algorithm using one or more processors, to determine assessed damage for the vehicle;
access actual damage information for the vehicle;
determine differences between the actual damage and the assessed damage; and
iterate the applicable machine learning algorithm based on the differences to improve its damage assessment accuracy.

### Claim 19 (depends on 18)

The storage media of claim 18, wherein the instructions, when executed by the one or more processors, cause the one or more processors to:
access second image data representing one or more digital images of one or more other vehicles with known actual damage; and
process the image data and the second image data, with the applicable machine learning algorithm using one or more processors, to determine the assessed damage.

### Claim 20 (depends on 19)

The storage media of claim 19, wherein the instructions, when executed by the one or more processors, cause the one or more processors to:
retrieve insurance claim data for the one or more other vehicles; and
process the image data, the second image data and the insurance claim data, with the applicable machine learning algorithm using one or more processors, to determine the assessed damage.