Image processing system for vehicle damage

• US 10,380,696 B1
• Filed: 03/24/2016
• Issued: 08/13/2019
• Est. Priority Date: 03/18/2014
• Status: Active Grant

• Current Assignee: CCC Intelligent Solutions, Inc.
• Original Assignee: CCC Information Services, Inc.
• Inventors: Haller, Jr., John Laurence, Smith, Darrin A.
• Primary Examiner(s): Akintola, Olabode
• Application Number:
  US15/079,356
• Time in Patent Office: 1,237 Days
• Field of Search: None
• CPC Class Codes: G06Q 30/0283 Price estimation or determination
• CPC Class Codes: G06Q 40/08 Insurance
• CPC Class Codes: G06V 20/20 in augmented reality scenes
• CPC Class Codes: G06V 2201/08 Detecting or categorising vehicles

Abstract

An image processing system automatically processes a plurality of images of a damaged vehicle and determines replacements parts that are needed to repair the vehicle therefrom. The system includes a network interface via which the images are received, an image extraction component that generates a set of image attributes from the received images, a parts identifier component that generates indications of the needed replacement parts based on the image attributes and an information identification model, and an output interface via which the generated indications are provided. Additionally, the system includes a data storage entity storing data from multiple historical vehicle insurance claims, and a model generation component that generates the information identification model based on the historical claim data. The information identification model includes independent variable(s) corresponding to a set of image attributes that are more strongly correlated to replacement parts than are other attributes of the historical claim data.

Claims

1. An image processing system, comprising:

one or more data storage devices comprising non-transitory, tangible computer-readable media storing thereon historical claim data of a plurality of historical vehicle insurance claims, the historical claim data including a plurality of claim attributes including image attribute data of the plurality of historical vehicle insurance claims, respective indications of actual replacement parts of the plurality of historical vehicle insurance claims, and a plurality of other claim attributes of the plurality of historical vehicle insurance claims;

an access mechanism to the historical claim data; and

one or more computing devices comprising;

a network interface via which a plurality of images of a damaged vehicle is received;

an image attribute extraction component configured to generate, by operating on the plurality of images of the damaged vehicle, a set of image attributes indicative of a content of at least some of the plurality of images of the damaged vehicle, the set of image attributes including one or more image attribute types and respective values of the one or more image attribute types;

a parts identifier component configured to generate, based on the set of image attributes indicative of the damaged vehicle, respective indications of one or more replacement parts needed to repair the damaged vehicle, the parts identifier component including an information identification model generated by accessing, via the access mechanism, the historical claim data of the plurality of historical vehicle insurance claims and performing a regression analysis on the accessed historical claim data to determine a subset of a plurality of image attributes that are more strongly correlated to actual replacement parts corresponding to the plurality of historical vehicle insurance claims than are other attributes of the plurality of claim attributes; and

an output interface via which the generated, respective indications of the one or more replacement parts are provided to at least one of the network interface or a user interface.

2. The image processing system of claim 1, wherein the plurality of historical vehicle insurance claims includes historical vehicle insurance claims corresponding to a plurality of insurance carriers.

3. The image processing system of claim 1, wherein the generated respective indications of the one or more replacement parts are provided via the output interface of the image processing system at First Notice of Loss (FNOL).

4. The image processing system of claim 1, wherein the set of image attributes comprises an image attribute vector including the respective values of the one or more image attribute types, and wherein the respective values of the one or more image attribute types comprise numerical values.

5. The image processing system of claim 1, wherein:

the image attribute extraction component comprises a feature detection component;

the one or more image attribute types include one or more of an edge, a corner, a blob, a ridge, a feature location, a feature size, a feature dimension, an intersection of features, a feature determined based on a scale-invariant feature transformation, a feature determined based on a threshold, a feature determined based on a template match, or another image attribute type; and

at least some of the respective values of the one or more image attribute types are generated by the feature detection component operating on the plurality of images of the damaged vehicle.

6. The image processing system of claim 1, wherein:

the image attribute extraction component comprises an image transformation component;

the one or more image attribute types include one or more of a scale corresponding to one or more images, an orientation of the damaged vehicle within one or more images, a texture, an edge, a corner, a point, a radius, a shape, a lighting orientation, a shadow orientation, a lighting location, a shadow location, or another image attribute type; and

at least some of the respective values of the one or more image attribute types are generated by the image transformation component at least one of transforming or filtering the plurality of images of the damaged vehicle.

7. The image processing system of claim 1, wherein:

the image attribute extraction component comprises a metric determination component;

the one or more image attribute types include one or more metrics from a set of metrics including a statistical measure, a gradient measure, and another metric; and

at least some of the respective values of the one or more image attribute types are calculated by the metric determination component operating on the plurality of images of the damaged vehicle.

8. The image processing system of claim 7, wherein the one or more metrics include at least one of:

a frequency of a particular attribute within the plurality of images,

an average of the particular attribute within the plurality of images,

a distribution of the particular attribute within the plurality of images,

another metric corresponding to the particular attribute within the plurality of images,

an average corresponding to a color, a texture, or a contrast property within the plurality of images,

a gradient corresponding to the color, the texture, or the contrast property within the plurality of images,

a frequency corresponding to the color, the texture, or the contrast property within the plurality of images,

a distribution corresponding to the color, the texture, or the contrast property within the plurality of images, or

another metric corresponding to the color, the texture, or the contrast property within the plurality of images.

9. The image processing system of claim 1, wherein:

at least one of the one or more image attribute types included in the set of image attributes corresponding to the damaged vehicle is an independent variable of the information identification model;

the respective values of the at least one of the one or more image attribute types included in the set of image attributes corresponding to the damaged vehicle are provided as an input to the information identification model; and

at least one of the respective indications of the one or more replacement parts needed to repair the damaged vehicle correspond to respective dependent variables of the information identification model and are received via an output of the information identification model.

10. The image processing system of claim 1, wherein the parts identifier component generates the respective indications of the one or more replacement parts needed to repair the damaged vehicle further based on at least one of a level of repair quality or a customer satisfaction score.

11. The image processing system of claim 1, wherein the respective indications of the one or more replacement parts needed to repair the damaged vehicle includes at least one indication of a replacement part type selected from:

OEM (Original Equipment Manufacturer), new, recycled, reconditioned, or another replacement part type.

12. The image processing system of claim 1, wherein the parts identifier component is configured to generate respective indications of one or more labor types and/or labor costs needed to repair the damaged vehicle based on the image attribute data of the plurality of historical vehicle insurance claims;

the regression analysis from which the information identification model is generated determines another subset of the plurality of claim attributes that are more strongly correlated to actual labor types and/or labor costs corresponding to the plurality of historical vehicle insurance claims than are other attributes of the plurality of claim attributes;

the historical claim data includes respective indications of the actual labor types and/or labor costs of the plurality of historical vehicle insurance claims; and

the generated, respective indications of the one or more labor types and/or labor costs are provided to the at least one of the network interface or the user interface via the output interface of the image processing system.

13. The image processing system of claim 1, wherein the parts identifier component is configured to generate a settlement amount corresponding to repairing the damaged vehicle based on the image attribute data of the plurality of historical vehicle insurance claims;

the regression analysis from which the information identification model is generated determines another subset of the plurality of claim attributes that are more strongly correlated to actual settlement amounts corresponding to the plurality of historical vehicle insurance claims than are other attributes of the plurality of claim attributes;

the historical claim data includes respective indications of the actual settlement amounts of the plurality of historical vehicle insurance claims; and

the generated settlement amount is provided to the at least one of the network interface or the user interface via the output interface of the image processing system.

14. The image processing system of claim 1, wherein the parts identifier component is configured to generate a payout amount of an insured party corresponding to repairing the damaged vehicle based on the image attribute data of the plurality of historical vehicle insurance claims;

the regression analysis from which the information identification model is generated determines another subset of the plurality of claim attributes that are more strongly correlated to actual payout amounts corresponding to the plurality of historical vehicle insurance claims than are other attributes of the plurality of claim attributes;

the historical claim data includes respective indications of the actual payout amounts of the plurality of historical vehicle insurance claims; and

the generated payout amount is provided to the at least one of the network interface or the user interface via the output interface of the image processing system.

15. The image processing system of claim 1, wherein the parts identifier component is configured to generate respective indications of one or more respective values of one or more other parameters corresponding to repairing the damaged vehicle based on the image attribute data of the plurality of historical vehicle insurance claims;

the regression analysis from which the information identification model is generated determines another subset of the plurality of claim attributes that are more strongly correlated to actual respective values of the one or more other parameters corresponding to the plurality of historical vehicle insurance claims than are other attributes of the plurality of claim attributes;

the historical claim data includes respective indications of the actual respective values of the one or more other parameters of the plurality of historical vehicle insurance claims; and

the generated, respective indications of the one or more respective values of the one or more other parameters are provided to the at least one of the network interface or the user interface via the output interface of the image processing system.

16. The image processing system of claim 15, wherein the one or more other parameters include at least one of:

a level of repair quality, a customer satisfaction score, a settlement amount, a towing cost, a paint cost, a tire cost, a re-inspection occurrence, a re-inspection cost, a hazardous waste disposal cost, a replacement part type, a payout amount for an insured party, a supplement occurrence, a supplement cost, a number of labor hours, a labor cost, or a labor type.

17. The image processing system of claim 1, wherein at least one of the network interface, the image attribute extraction component, the parts identifier component, or the output interface comprises hardware.

18. The image processing system of claim 17, wherein at least one of the network interface, the image attribute extraction component, the parts identifier component, or the output interface comprises software or firmware.

19. The image processing system of claim 18, wherein the software comprises a set of computer-executable instructions stored on one or more memories, the hardware comprises the one or more memories, and the image processing system further comprises one or more processors for executing the set of computer-executable instructions.

20. The image processing system of claim 1, wherein at least one of (i) the set of image attributes corresponding to the damaged vehicle, (ii) the respective indications of the one or more replacement parts needed to repair the damaged vehicle, or (iii) respective indications of one or more values of one or more other parameters corresponding to repairing the damaged vehicle and generated by the parts identifier component are at least one of (a) stored in at least one of the one or more data storage devices or in another one or more data storage devices, or (b) provided to the at least one of the network interface or the user interface via the output interface of the image processing system.

21. A method of image processing, comprising:

obtaining, via a network interface of an image processing system, a plurality of images of a damaged vehicle;

generating, by using an image attribute extraction component of the image processing system operating on the plurality of images of the damaged vehicle, a set of image attributes indicative of a content of at least some of the plurality of images of the damaged vehicle, the set of image attributes including one or more image attribute types and respective values of the one or more image attribute types;

generating, by using a parts identifier component of the image processing system, and based on the set of image attributes indicative of the damaged vehicle, one or more indications of one or more replacement parts needed to repair the damaged vehicle, the parts identifier component including an information identification model generated by performing a regression analysis on historical claim data of a plurality of historical vehicle insurance claims,

the historical claim data including a plurality of claim attributes including image attribute data of the plurality of historical vehicle insurance claims, respective indications of actual replacement parts of the plurality of historical vehicle insurance claims, and a plurality of other claim attributes of the plurality of historical vehicle insurance claims, and

the regression analysis to determine a subset of a plurality of image attributes that are more strongly correlated to actual replacement parts corresponding to the plurality of historical vehicle insurance claims than are other attributes of the plurality of claim attributes; and

providing, via an output interface of the image processing system to a recipient, the generated, one or more indications of the one or more replacement parts needed to repair the damaged vehicle.

22. The method of claim 21, wherein the method is performed at First Notice of Loss (FNOL) corresponding to the damaged vehicle.

23. The method of claim 21, further comprising at least one of ordering the one or more replacement parts, or repairing the damaged vehicle using the one or more replacement parts.

24. The method of claim 21, wherein generating the set of image attributes indicative of the content of at least some of the plurality of images of the damaged vehicle comprises at least one of:

detecting one or more features within the plurality of images, transforming at least one image of the plurality of images, filtering the at least one image or another image of the plurality of images, or calculating one or more metrics for at least one feature, color, texture, or contrast property within the plurality of images.