Classification model generation method and device and medical image classification method and device

Embodiments of the application disclose a classification model generation method and device and a medical image classification method and device. The classification model generation method includes the steps of obtaining an original medical image, extracting at least one image feature of the original medical image and/or generating at least one transformed image corresponding to the original medical image to form groups of training data respectively, generating a basic classification model through training of each group of training data, and forming a common medical image classification modeltogether by all the basic classification models, wherein the generated medical image classification model can classify medical images, and classification results eliminate the influence of subjectivity and are more accurate. Meanwhile, the embodiments of the application can construct a medical image classification model for any type of medical images, and the mode of constructing the medical imageclassification model has