METHOD, DEVICE, AND SYSTEM FOR IMAGE RECOGNITION MODEL TRAINING AND IMAGE RECOGNITION

摘要

The present application relates to the technical field of computers and specifically relates to a method, device, and system for image recognition model training and image recognition. The recognition method: acquiring an image to be recognized; extracting image feature information of said image; acquiring a disease type recognition result of said image on the basis of a preset image recognition model and with the image feature information of said image serving as an input parameter, where the image recognition model is trained by employing a training image sample set at least comprising strong tag training image samples to determine the disease type recognition resu the strong tag training image samples express image samples having strong tag information, the strong tag information at least comprises label information of disease types and lesions. Image feature information of a certain disease type can be accurately positioned on the basis of the lesions, thus reducing noise and increasing reliability and accuracy.