Attention pooling-based end-to-end histopathological image classification method

The invention discloses an attention pooling-based end-to-end histopathological image classification method, which comprises the specific steps of S1, segmenting a histopathological image into sliceswith specified sizes, removing excessive slices in a background area, and forming a packet by the remaining slices; S2, taking the packet obtained in the step S1 as input, and training a deep neural network by using a standard multi-example learning method. S3, scoring all slices by using the trained deep neural network, taking m slices with the highest and lowest scores of each full-slide image,and gathering the m slices into a new packet. S4, establishing a deep neural network containing an attention module, and training the network by using the new packet obtained in the step S3. S5, afterthe histopathological images to be classified are processed in the steps S1 and S3, using the model obtained in the step S4 to carry out classification. According to the method, a good classificationeffect can be obtained unde