Non-small cell lung cancer pathological section identification method based on deep convolutional neural network

Abstract

The invention discloses a non-small cell lung cancer pathological section identification method based on a deep convolutional neural network. The method comprises the following steps: acquiring pathological sections of non-small cell lung cancer in a public data set from TCGA; constructing a deep learning model for training; inputting the training data set into a convolutional neural network for training to obtain a learned convolutional neural network model; and inputting the training data set into a convolutional neural network for training to obtain a learned convolutional neural network model. According to the method, the Inception-v3 model and the CBAM attention mechanism are fused together, so that the classification of the non-small cell lung cancer is realized, and the network precision is improved through the attention mechanism; meanwhile, a deep convolutional neural network Inception-v3 experimental result shows that the non-small cell lung cancer pathological section identification method based on deep learning provided by the invention can effectively classify lung adenocarcinoma and lung squamous cell carcinoma, reduces the burden of doctors to a certain extent, and realizes very good performance in the field of medical image identification.

Description

technical field [0001] The invention belongs to the technical fields of medical imaging, image processing and deep learning, and in particular relates to a method for identifying pathological slices of non-small cell lung cancer based on a deep convolutional neural network. Background technique [0002] With the development of medical imaging and computer recognition technology, more and more medical inspection methods can present the inspection results in the form of digital images. As a result, a large amount of medical imaging data, such as pathological slides, CT (Computed Tomography), and MRI (Magnetic Resonance Imaging), etc. have also been generated. Digitally characterizing microscopic pathologies is a new field of pathology. Access to digital images facilitates remote primary diagnostic work, remote consultation and balance of work for physicians. [0003] Due to the development of deep learning, researchers have trained various deep architectures using large imag...

AI Technical Summary

Problems solved by technology

Traditional methods of diagnosis rely on experienced pathologists for confirmation, which largely relies on the experience and knowledge of doctors

Images