Deep learning image processing method for determining vehicle damage

A tangible, non-transitory computer-readable medium storing instructions that, when executed by one or more processors, cause the one or more processors to:

• access a historical training dataset populated with digital images of damaged vehicles and historical claim data related to damaged vehicles;
• generate a predictive similarity model by using the historical training dataset to train a deep learning neural network,
• wherein training the deep learning neural network includes learning:
• damaged vehicle image characteristics that are predictive of one or more of damage level, repair time, or repair cost, and
• a degree to which each of the damaged vehicle image characteristics is predictive of one or more of damage level, repair time, or repair cost;
• determine, using the predictive similarity model, a similarity score for a given two digital images of damaged vehicles;
• determine, based on the similarity score, that the given two digital images satisfy one or more similarity criteria;
• in response to a determination that the two digital images satisfy the one or more similarity criteria, cause the one or more processors to do one or more of the following:
• display the two digital images on a single page;
• and
• predict a damage level, repair time, or repair cost for a first vehicle of the damaged vehicles depicted in the given two digital images based on a damage level, repair time, or repair cost for a second vehicle of the damaged vehicles depicted in the given two digital images.

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
• access a historical training dataset populated with digital images of damaged vehicles and historical claim data related to damaged vehicles;
• generate a predictive similarity model by using the historical training dataset to train a deep learning neural network, wherein training the deep learning neural network includes learning:
• damaged vehicle image characteristics that are predictive of one or more of damage level, repair time, or repair cost, and
• a degree to which each of the damaged vehicle image characteristics is predictive of one or more of damage level, repair time, or repair cost;
• determine, using the predictive similarity model, a similarity score for a given two digital images of damaged vehicles;
• determine, based on the similarity score, that the given two digital images
• satisfy one or more similarity criteria;
• in response to a determination that the two digital images satisfy the one or more similarity criteria, cause the one or more processors to do one or more of the following:
• display the two digital images on a single page;
• and
• predict a damage level, repair time, or repair cost for a first vehicle of the damaged vehicles depicted in the given two digital images based on a damage level, repair time, or repair cost for a second vehicle of the damaged vehicles depicted in the given two digital images.