Method for generating traditional Chinese medicine prescription based on deep learning

The invention discloses a method for generating a traditional Chinese medicine prescription based on deep learning. The method comprises the steps of segmenting the face image of a patient to images in three granularities: organ granularity, local area granularity and face granularity; dividing a dataset which is composed of a pixel matrix and a multi-label vector into a training dataset and a testing dataset; constructing a convolutional neural network model based on a three-granularity face; training the training dataset, performing training of iterating an adjusting parameter by multiple times for obtaining a neural network model with excellent performance; and inputting the pixel matrix of the testing dataset into the neural network model for generating the traditional Chinese medicineprescription. According to the method of the invention, prescription making experience of the traditional Chinese medicine can be learned from data; and the traditional Chinese medicine prescriptionwhich is required by the p