Thoracic cavity X-ray diagnosis report generation method based on memory enhancement Transformer

The invention discloses a chest X-ray diagnosis report generation method based on memory enhancement Transform, which comprises the following steps: collecting a plurality of chest X-ray images and corresponding diagnosis reports according to actual needs, constructing word lists based on the diagnosis reports, and obtaining a word probability matrix and a word vector matrix of words contained in the word lists in each diagnosis report; forming a training sample together with the thoracic X-ray image to obtain a training sample set; constructing a diagnosis report word prediction model based on a Transform and a memory module; training the diagnosis report word prediction model by adopting the training sample, inputting a to-be-generated diagnosis report thoracic cavity X-ray image into the trained diagnosis report word prediction model to obtain a word probability matrix, and finally generating a final diagnosis report text by adopting a text sequence generation algorithm. According to the method, the memory