Deep learning image processing method for determining vehicle damage

Patent metadata

Field	Value
Publication number	US-2024289891-A1
Application number	US-202418658330-A
Country	US
Kind code	A1
Filing date	May 8, 2024
Priority date	Jun 29, 2017
Publication date	Aug 29, 2024
Grant date	—

Abstract

In a computer-implemented method and associated tangible non-transitory computer-readable medium, an image of a damaged vehicle may be analyzed to generate a repair estimate. A dataset populated with digital images of damaged vehicles and associated claim data may be used to train a deep learning neural network to learn damaged vehicle image characteristics that are predictive of claim data characteristics, and a predictive similarity model may be generated. Using the predictive similarity model, one or more similarity scores may be generated for a digital image of a newly damaged vehicle, indicating its similarity to one or more digital images of damaged vehicles with known damage level, repair time, and/or repair cost. A repair estimate may be generated for the newly damaged vehicle based on the claim data associated with images that are most similar to the image of the newly damaged vehicle.

First claim

What is claimed:

1 . 1 . A computer-implemented method, comprising:
receiving a first digital image of a first damaged vehicle; determining one or more characteristics of the first digital image; identifying, using an image search engine, from a database of digital images of damaged vehicles, one or more digital images of one or more additional damaged vehicles, having one or more same characteristics as the first digital image; and predicting one or more of a damage level, a repair time, or a repair cost for the first damaged vehicle based on one or more of a known damage level, a known repair time, or a known repair cost for the one or more digital images of the one or more additional damaged vehicles.

2 . The computer-implemented method of claim 1 , further comprising:
receiving input from a user including a selection of one of the one or more additional damaged vehicles; and wherein predicting the one or more of the damage level, the repair time, or the repair cost for the first damaged vehicle based on one or more of a known damage level, a known repair time, or a known repair cost for the selected one or more additional damaged vehicles.

3 . The computer-implemented method of claim 1 , wherein predicting one or more of the damage level, the repair time, or the repair cost for the first damaged vehicle is further based on weighing one or more of the known damage level, the known repair time, or the known repair time for the one the one or more digital images of the one or more additional damaged vehicles based on a degree to which each of the one or more same characteristics of the respective digital images are predictive of damage level, repair time, or repair cost.

4 . The computer-implemented method of claim 1 , wherein identifying the one or more digital images of the one or more additional damaged vehicles using the image search engine is based on a similarity score associated with each of the one or more digital images when compared to the first digital image of the first damaged vehicle.

5 . The computer-implemented method of claim 4 , wherein identifying the one or more digital images of the one or more additional damaged vehicles using the image search engine is based on the similarity score associated with each of the one or more digital images, when compared to the first digital image of the first damaged vehicle, exceeding a similarity score threshold.

6 . The computer-implemented method of claim 4 , further comprising:
determining, using a predictive similarity model, the similarity score for the first digital image of the first damaged vehicle and another image from the one or more digital images of the one or more additional damaged vehicles, wherein the predictive similarity model is trained using a historical training dataset populated with digital images of damaged vehicles and historical claim data related to the damaged vehicles, to identify damaged vehicle image characteristics that are predictive of one or more of damage level, repair time, or repair cost.

7 . The computer-implemented method of claim 1 , further comprising: displaying the one or more digital images of one or more additional damaged vehicles.

8 . A tangible, non-transitory computer-readable medium storing instructions that, when executed by one or more processors, cause the one or more processors to perform actions comprising:
receiving a first digital image of a first damaged vehicle; determining one or more characteristics of the first digital image; identifying, using an image search engine, from a database of digital images of damaged vehicles, one or more digital images of one or more additional damaged vehicles, having one or more same characteristics as the first digital image; and predicting one or more of a damage level, a repair time, or a repair cost for the first damaged vehicle based on one or more of a known damage level, a known repair time, or a known repair cost for the one or more digital images of the one or more additional damaged vehicles.

9 . The tangible, non-transitory computer-readable medium of claim 8 , the actions further comprising:
receiving input from a user including a selection of one of the one or more additional damaged vehicles; and wherein predicting the one or more of the damage level, the repair time, or the repair cost for the first damaged vehicle based on one or more of a known damage level, a known repair time, or a known repair cost for the selected one or more additional damaged vehicles.

10 . The tangible, non-transitory computer-readable medium of claim 8 , wherein predicting one or more of the damage level, the repair time, or the repair cost for the first damaged vehicle is further based on weighing one or more of the known damage level, the known repair time, or the known repair time for the one the one or more digital images of the one or more additional damaged vehicles based on a degree to which each of the one or more same characteristics of the respective digital images are predictive of damage level, repair time, or repair cost.

11 . The tangible, non-transitory computer-readable medium of claim 8 , wherein identifying the one or more digital images of the one or more additional damaged vehicles using the image search engine is based on a similarity score associated with each of the one or more digital images when compared to the first digital image of the first damaged vehicle.

12 . The tangible, non-transitory computer-readable medium of claim 11 , wherein identifying the one or more digital images of the one or more additional damaged vehicles using the image search engine is based on the similarity score associated with each of the one or more digital images, when compared to the first digital image of the first damaged vehicle, exceeding a similarity score threshold.

13 . The tangible, non-transitory computer-readable medium of claim 11 , the actions further comprising:
determining, using a predictive similarity model, the similarity score for the first digital image of the first damaged vehicle and another image from the one or more digital images of the one or more additional damaged vehicles, wherein the predictive similarity model is trained using a historical training dataset populated with digital images of damaged vehicles and historical claim data related to the damaged vehicles, to identify damaged vehicle image characteristics that are predictive of one or more of damage level, repair time, or repair cost.

14 . The tangible, non-transitory computer-readable medium of claim 8 , the actions further comprising:
displaying the one or more digital images of one or more additional damaged vehicles.

15 . A computer system comprising:
one or more processors; and a memory storing instructions that, when executed by the one or more processors, cause the one or more processors to perform actions comprising:
receiving a first digital image of a first damaged vehicle;
determining one or more characteristics of the first digital image;
identifying, using an image search engine, from a database of digital images of damaged vehicles, one or more digital images of one or more additional damaged vehicles, having one or more same characteristics as the first digital image; and
predicting one or more of a damage level, a repair time, or a repair cost for the first damaged vehicle based on one or more of a known damage level, a known repair time, or a known repair cost for the one or more digital images of the one or more additional damaged vehicles.

16 . The computer system of claim 15 , the actions fur

Assignees

• State Farm Mutual Automobile Insurance Co

Inventors

• Yang He
• Sliz Bradley A
• Clymer Carlee A
• Andrus Jennifer Malia

Classifications

• G06N3/08

  Learning methods · CPC title

• G06N3/0464

  Convolutional networks [CNN, ConvNet] · CPC title

• G06N3/09

  Supervised learning · CPC title

• G06V10/82

  using neural networks · CPC title

• G06V10/761

  Proximity, similarity or dissimilarity measures · CPC title