Training methods, devices, equipment, and media for skin image classification models

Abstract

This application discloses a training method, apparatus, medium, and device for a skin image classification model. The method includes: acquiring a training set, wherein the training set includes one or more skin images and corresponding patient information; inputting the skin images into a skin image classification model, wherein the skin image classification model includes multiple learning tasks, one of which is a target task; adjusting the parameters of the skin image classification model based on the first output information of the multiple learning tasks and the patient information; inputting the skin images into the skin image classification model, and adjusting the parameters based on the second output information of the target task and the patient information. This application's method solves the problem of low accuracy in skin image classification models and improves the quality of online medical services.

Images