A PATHOLOGICAL DIAGNOSIS METHOD FOR GRADING PANCREATIC DUCTAL ADENOCARCINOMAS BASED ON MORPHOLOGICAL FEATURES

摘要

The present invention relates to a method to diagnose a pancreatic ductal adenocarcinomas grade based on morphological features which extracts morphological features from tissue cell images and diagnoses pancreatic ductal adenocarcinomas based on the extracted features. The present invention comprises: a first step of extracting morphological features from the tissue cell images; a second step of forming a diagnosis model using a classifier based on supervised learning; and a third step of diagnosing pancreatic ductal adenocarcinomas of the tissue cell image using the diagnosis model, wherein the diagnosis model prepares a composition determining whether to be pancreatic ductal adenocarcinomas or a grade about the inputted morphological features. Through the method to diagnose a pancreatic ductal adenocarcinomas grade, morphological features can be extracted and pancreatic ductal adenocarcinomas can precisely be distinguished; and each classifier can be classified by learning the diagnosis model using the extracted features.