Prediction method, device and equipment for super-resolution gene expression spectrum

The invention relates to the technical field of neural network prediction, and discloses a method, a device and equipment for predicting a super-resolution gene expression map. The method comprises the following steps: dividing a histological image into a large image block and a small image block according to the pixel size, inputting the large image block, the small image block and an original histological image into a pre-training hierarchical feature extraction model, and outputting a hierarchical histological feature map; and taking each super-pixel point in the hierarchical histological feature map as a node, dividing the hierarchical histological feature map into a series of sub-maps, inputting the series of sub-maps into the trained super-resolution gene expression prediction model, and outputting a predicted super-resolution gene expression map. According to the method, the ViT module is used for carrying out multi-scale extraction on histological image features, and complex association information be