Deep learning-based lymphoma pathological image intelligent identification method

Abstract

The invention discloses a deep learning-based lymphoma pathological image intelligent identification method. The method comprises the steps of preprocessing lymphoma pathological section image data; constructing a full convolutional neural network for segmenting a lymphatic tissue region, wherein the full convolutional neural network comprises an encoder sub-network and a decoder sub-network; constructing a lymphoma three-classification convolutional neural network under high resolution, wherein the lymphoma three-classification convolutional neural network is composed of six convolutional layers and three full-connection layers which are connected in sequence; and training the full convolutional neural network and the lymphoma three-classification convolutional neural network to finally obtain a lymphoma pathological section image classification model, and sequentially passing through the full convolutional neural network and the lymphoma three-classification convolutional neural network during testing to finally obtain a lymphoma classification result. Reliable intermediate data are provided for pathologists to judge lymphoma subtype categories, and auxiliary diagnosis referenceis provided for the pathologists to classify lymphoma subtypes by analyzing digitally scanned lymphoma pathological images, so that the pathologists are helped to quickly judge lymphoma conditions ofpatients.

Description

technical field [0001] The invention relates to an auxiliary diagnosis system for lymphoma. In particular, it relates to an intelligent recognition method for lymphoma pathological images based on deep learning. Background technique [0002] Lymphoma is one of the common malignant tumors in my country. Due to its complex and diverse pathological types and no specific histopathological manifestations, the clinicopathological diagnosis is easily confused with other tumors and easily leads to misdiagnosis. It is currently a relatively difficult clinical pathological diagnosis. class of tumors. The pathological misdiagnosis rate of lymphoma is 10% to 33.33%. Misdiagnosed patients cannot receive timely treatment and often miss the best treatment opportunity, which seriously affects the treatment and prognosis of patients. Among them, diffuse large B-cell lymphoma (DLBCL) is the most common non-Hodgkin's lymphoma (NHL), accounting for about 30% to 40% of all NHLs. It is a group o...

AI Technical Summary

Problems solved by technology

However, due to the complexity of histopathological manifestations of lymphoma and the high-resolution characteristics of pathological images, computer-aided diagnosis is difficult

Images