Method for generating a license plate defacement classification model, license plate defacement classification method, electronic device and storage medium

A license plate defacement classification method, which is based on a license plate defacement classification model generated by using operations of:

• acquiring training data, wherein the training data comprises a plurality of annotated vehicle images, annotated content comprises information indicating that a license plate is defaced or is not defaced, and the annotated content further comprises location information of a license plate area;
• and
• training a first neural network by using the training data, to obtain the license plate defacement classification model for predicting whether the license plate in a target vehicle image is defaced,
• wherein loss functions used in the training comprise a cross-entropy loss function and a class activation map loss function, the cross-entropy loss function is used for performing supervised learning on whether the license plate is defaced, and the class activation map loss function is used for performing supervised learning on a license plate location, wherein the class activation map loss function is obtained by comparing the annotated vehicle image with a class activation map corresponding to the annotated vehicle image,
• wherein the license plate defacement classification method comprises:
• inputting a target vehicle image to the license plate defacement classification model;
• predicting, by the license plate defacement classification model, whether a license plate in the target vehicle image is defaced;
• and
• receiving a prediction result output by the license plate defacement classification model, wherein the prediction result comprises information indicating that the license plate is defaced or is not defaced.

An electronic device, comprising:

• at least one processor;
• and
• a memory communicatively connected with the at least one processor,
• wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to execute operations of:
• acquiring training data, wherein the training data comprises a plurality of annotated vehicle images, annotated content comprises information indicating that a license plate is defaced or is not defaced, and the annotated content further comprises location information of a license plate area;
• and
• training a first neural network by using the training data, to obtain the license plate defacement classification model for predicting whether the license plate in a target vehicle image is defaced,
• wherein loss functions used in the training comprise a cross-entropy loss function and a class activation map loss function, the cross-entropy loss function is used for performing supervised learning on whether the license plate is defaced, and the class activation map loss function is used for performing supervised learning on a license plate location, wherein the class activation map loss function is obtained by comparing the annotated vehicle image with a class activation map corresponding to the annotated vehicle image.