Method and device for evaluating vehicle damage recognition model

Abstract

The embodiment of this specification provides a method and device for evaluating a car damage recognition model. The method includes firstly obtaining a test sample, the test sample corresponds to a car damage picture and multiple sets of labeling data, the multiple sets of labeling data are at least based on multiple labeling personnel The car damage picture is generated by labeling; then determine the intersection and union of multiple sets of labeled data, and determine the set of significant damage objects in the test sample according to the intersection; and determine the non-destructive objects in the test sample according to the difference between the intersection and the union. A collection of significantly damaged objects. In addition, the above-mentioned car damage pictures are input into the pre-trained car damage recognition model to obtain a set of predicted damage objects output by the model for the test sample. Therefore, the test result of the vehicle damage recognition model for the test sample can be determined according to the relationship between the predicted damage object set and the salient damage object set and the non-notice damage object set.

Description

technical field [0001] One or more embodiments of this specification relate to the field of machine learning, and in particular to a method and device for evaluating a vehicle damage recognition model. Background technique [0002] In the process of traditional auto insurance claims, the insurance company needs to send professional survey and loss assessment personnel to the accident scene to conduct on-site survey and assessment of damage, provide the vehicle maintenance plan and compensation amount, and take photos of the scene, and keep the damage assessment photos for background verification Personnel check damage and price. Due to the need for manual damage assessment, insurance companies need to invest a lot of labor costs and professional knowledge training costs. From the experience of ordinary users, the claim settlement process takes as long as 1-3 days due to waiting for the manual surveyor to take pictures on site, the damage assessor to determine the damage at ...

AI Technical Summary

Problems solved by technology

Due to the need for manual survey and loss assessment, insurance companies need to invest a lot of labor costs and professional knowledge training costs

From the experience of ordinary users, the claim settlement process is as long as 1-3 days due to waiting for the manual surveyor to take pictures on the spot, the damage assessor to determine the damage at the repair location, and the loss checker to verify the damage in the background. , poor experience

[0003] Aiming at the industry pain points of huge labor costs in traditional auto insurance claims, some intelligent image loss determination solutions are proposed

Images