Medical image classification method based on feature fusion and attention mechanism

The invention discloses a medical image classification method based on feature fusion and an attention mechanism. The method comprises the steps of image preprocessing; and feature extraction based on a convolutional neural network backbone network. Carrying out multi-scale feature fusion; interacting the attention pyramid, screening out discriminative information, and focusing the discriminative information on the important information; and outputting a classification result through an average pooling layer. According to the classification method, features are extracted in a convolution mode, the number of parameters is small, the operation speed is high, the method is suitable for features of a small number of samples of medical image data, and excessive weight parameters do not need to be updated; meanwhile, experiments prove that the method can well adapt to some difficult-to-classify medical image data sets with large inter-class differences and small class differences. The method is not only suitable fo