Pancreatic cancer pathological image classification method and system based on deep learning

Abstract

The invention provides a pancreatic cancer pathological image classification method and system based on deep learning. The method comprises the following steps: carrying out image preprocessing; constructing an image screening network model; constructing a pancreatic cancer classification network model; cascading the image screening network model and the pancreatic cancer classification network model to construct an end-to-end pancreatic cancer rapid on-site evaluation system; and judging and classifying pancreatic cancer pathological images. According to the method, a transfer learning technology and an attention mechanism are applied, and two different neural networks are cascaded, so that a classification system with the precision equivalent to that of a pathologist can be obtained, and a real-time and accurate pancreatic cancer rapid on-site evaluation tool is provided for clinicians.

Description

technical field [0001] The invention relates to the field of artificial intelligence-assisted pathological image recognition, in particular to a method and system for classifying pancreatic cancer pathological images based on deep learning. Background technique [0002] Pancreatic cancer is a highly malignant tumor of the digestive system. Its early diagnosis rate is low and its prognosis is extremely poor. It is one of the most malignant tumors. In recent years, the incidence of pancreatic cancer has increased significantly both at home and abroad. The latest statistics show that the incidence of pancreatic cancer in my country ranks ninth among malignant tumors, and its mortality rate reaches sixth, and the mortality rate in Western countries ranks fourth. bit. Due to the high degree of malignancy of pancreatic cancer, even if patients can be treated surgically, the five-year survival rate is only 6%. Therefore, early and precise diagnosis and treatment of pancreatic cance...

AI Technical Summary

Problems solved by technology

Aiming at the problem of less data and more background and noise areas in pancreatic pathological images, the present invention first uses transfer learning technology to transfer neural network parameters that have been pre-trained and well-performed in other data sets and then trained in this data set to solve the problem of The overfitting problem caused by the small amount of data can improve the generalization ability of the model

Images