Paint blending determination
The present invention relates to assessing the damage and repairs needed to damaged vehicles. More particularly, the present invention relates to assessing vehicle damage using primarily photos of damaged vehicles and information provided by drivers or insurers, to determine whether vehicle body parts to be replaced or repaired require paint blending. Aspects and/or embodiments seek to provide a method and system to determine whether a part of a damaged vehicle requires paint blending.
1. A computer-implemented method of determining paint blending requirements for a part of a damaged vehicle, comprising:
receiving one or more images of the part of the damaged vehicle;
receiving a plurality of items of metadata, the items of metadata comprising:
a score indicative of a repair or replace value for the part;
one or more scores indicative of a damage value of one or more neighbouring parts; and
one or more vehicle properties;
determining, using (i) a trained model, (ii) the one or more images of the part of the damaged vehicle and (iii) the plurality of items of metadata, a paint blending value for the part indicating whether a paint blending operation is to be performed with any of the one or more neighbouring parts, wherein the trained model corresponds to a region or the part of the damaged vehicle; and
outputting the determined paint blending value for the part.
2. The method of claim 1 , wherein, when determining the paint blending value for the part using a trained model, each item of metadata is provided in a sequential order.
3. The method of claim 1 , wherein the trained model comprises a plurality of sequential layers, and the one or more images of the part of the damaged vehicle are provided to the first of the plurality of sequential layers and each item of metadata is provided to one of the subsequent plurality of sequential layers.
4. The method of claim 1 , wherein the one of more vehicle properties comprises any or any combination of: a value indicating a type of the vehicle; a number of doors of the vehicle; and/or a value indicating a color of the vehicle.
5. The method of claim 1 , further comprising:
determining a paint color for use in any paint blending operation represented by the determined paint blending value for the part.
6. The method of claim 1 , wherein a plurality of images of the part of the damaged vehicle are received and a plurality of paint blending values for the part are determined, the method further comprising:
determining a pooled paint blending value for the part from the plurality of paint blending values for the part.
7. The method of claim 1 , wherein the trained model comprises any one or any combination of: a neural network; a convolutional neural network; and/or a recurrent neural network.
8. The method of claim 1 , wherein the score indicative of the repair or replace value for the part is a vector representing one or more quantitative values indicating the level of damage to the part.
9. The method of claim 1 , wherein the one or more scores indicative of the damage value of one or more neighboring parts are each vectors representing one or more quantitative values indicating a level of damage to each neighboring part.
10. The method of claim 1 , further comprising:
determining a plurality of parts of the damaged vehicle that are represented in the plurality of images of the damaged vehicle comprising the use of a plurality of classifiers.
11. The method of claim 10 , wherein each one of the plurality of classifiers is operable to detect each of the parts of the damaged vehicle.
12. The method of claim 1 , further comprising:
determining the relevant images showing the part.
13. The method of claim 1 , wherein the part of the damaged vehicle comprises any or any combination of: normalized parts of the vehicle; normalized regions of the vehicle; normalized zones of the vehicle; standardized parts of the vehicle; standardized regions of the vehicle; standardized zones of the vehicle.
14. A non-transitory computer-readable medium comprising computer-executable instructions which, when executed, perform the method of claim 1 .