METHOD AND APPARATUS FOR DETERMINING SEVERITY OF DAMAGE ON TARGET OBJECT, ELECTRONIC DEVICE, AND STORAGE MEDIUM

摘要

A method for determining the severity of damage on a target object, relating to the field of artificial intelligence, and comprising: by means of tags, classifying an acquired original image training set to obtain a standard image training set (S1), constructing a target object damage severity classification model comprising a classification sub-network and an edge regression sub-network (S2), using the classification sub-network to generate an anchor edge and a real edge for each image in the standard image training set, and by means of the edge regression sub-network, performing edge regression on the anchor edges and the real edges to obtain a predicted loss value set, and on the basis of the predicted loss value set, obtaining a standard target object damage severity classification model (S3), and using the target object damage severity classification model to analyze a damage image set of a target object to obtain a target object damage severity identification result (S4) The invention further relates to blockchain technology, and the data required for training the model can be stored in nodes of a blockchain. The precision of recognizing the severity of damage on a target object can thus be improved.