Intelligent vehicle repair estimation system

A system, comprising:

• a database storing a listing of possible parts for repairing vehicles and an indication of a respective jurisdiction corresponding to each possible part;
• one or more memories storing a plurality of routines including a parts prediction routine and an estimate generation routine;
• one or more interface modules;
• and
• one or more processors interfacing with the database, the one or more memories, and the one or more interface modules, the one or more processors configured to execute the plurality of routines to cause the system to:
• train a machine learning parts prediction model to determine a set of image attributes that are more closely correlated to vehicle parts that are needed to repair vehicle damage as compared to other image attributes, the training of the machine learning parts prediction model based on first historical data associated with repairing damaged vehicles, the first historical data including image attributes of historical images of damaged vehicles and global identifiers of parts that were utilized to repair the damaged vehicles depicted within the historical images, and the global identifiers categorically identifying respective vehicle parts across a plurality of jurisdictions;
• train a machine learning context mapping model to predict, for each jurisdiction of the plurality of jurisdictions, respective estimations of labor operations needed to repair damaged vehicles by using one or more jurisdictionally-based vehicle parts, the training of the machine learning context mapping model based on a set of jurisdictional requirements of the each jurisdiction and second historical data associated with repairing damaged vehicles, the second historical data indicating historical labor operations that were utilized for repairing damaged vehicles by replacing and/or repairing respective jurisdictionally-based vehicle parts;
• obtain, via the one or more interface modules, a set of image attributes of one or more images of a damaged vehicle, the set of image attributes indicative of respective locations and content of damages to the vehicle depicted within the one or more images;
• process, via the parts prediction routine and by utilizing the trained machine learning parts prediction model, the set of image attributes to obtain respective global identifiers of an initial set of parts associated with damages to the damaged vehicle and, for each part of at least some of the initial set of parts, at least one of:
• a respective likelihood of part replacement or a respective likelihood of part repair for repairing the damaged vehicle;
• determine, via the estimate generation routine and based on a jurisdiction corresponding to the damaged vehicle, the respective global identifiers of the initial set of parts, the part replacement likelihoods, the part repair likelihoods, and the listing of possible parts, a jurisdictionally-based set of parts corresponding to the damaged vehicle;
• predict, via the estimate generation routine and by utilizing the trained machine learning context mapping model, (i) for each part of at least some of the jurisdictionally-based set of parts, a respective indication as to whether each jurisdictionally-based part is to be repaired or replaced, and (ii) a set of labor operations needed to repair the damaged vehicle;
• and
• transmit, via the one or more interface modules to a recipient, an estimate for repairing the damaged vehicle, the estimate indicating the set of jurisdictionally-based set of parts and the set of labor operations.