Skin disease image classification method based on three-branch feature fusion network

The invention discloses a dermatoscope image classification method based on a three-branch feature fusion network. The method comprises the following steps: acquiring dermatoscope image data to be classified; inputting the dermatoscope image data to be classified into a three-branch layered multi-scale feature fusion network to obtain global features and multi-scale local features of each stage; fusing the global and local features of each stage through an adaptive hierarchical feature fusion module to obtain adaptive fusion feature information; the fusion information of the last stage is classified, and a classification result is obtained and output. Expanded image data are generated through a classifier gradient-guided diffusion model and used for training the three-branch hierarchical multi-scale feature fusion network, and a self-adaptive hierarchical feature fusion module of the three-branch feature fusion network is used for fully fusing and mining local and global features in dermatoscope images. And t