Auxiliary identification system and method for lung adenocarcinoma subtypes

Abstract

The invention discloses an auxiliary identification system and method for lung adenocarcinoma subtypes, and relates to a neural network. The system comprises an acquisition module for acquiring a digital pathological image and labeling to obtain a labeled image; wherein the features include lesion areas and real pathological image features; the processing module is used for processing the annotated image to obtain a processed image and storing the processed image into a database; the classification module divides the processed images into a training set, a verification set and a test set; thetraining module is used for training the training set to obtain an auxiliary identification model of the lung adenocarcinoma subtype; the verification module is used for inputting the verification setinto the auxiliary identification model for optimization to obtain an optimization model; the test module is used for inputting the test set into the optimization model to obtain a result and obtaintest accuracy; the comparison module is used for comparing the test accuracy with a test threshold value and retraining when the test accuracy is smaller than the test threshold value; and if not, storing the auxiliary identification optimization model. The kit has the beneficial effect of assisting a clinician in identifying the lung adenocarcinoma subtype.

Description

technical field [0001] The invention relates to the field of neural networks, in particular to an auxiliary identification system and method for subtypes of lung adenocarcinoma. Background technique [0002] Lung cancer is one of the most common malignant tumors in the world. The morbidity and mortality rate of lung cancer in my country rank first among malignant tumors, which seriously threatens the health of the people. Lung adenocarcinoma is the most common type of primary lung cancer, accounting for about half of all lung cancers. Histological classification of lung adenocarcinoma remains difficult because of tumor heterogeneity. In this case, the World Health Organization officially published the histological classification of lung cancer in 2015, and the most changed part is the histological classification of lung adenocarcinoma. According to the new classification criteria, the original simple BAC was replaced by adenocarcinoma in situ (AIS), with squamous growth a...

AI Technical Summary

Problems solved by technology

However, due to different algorithm models, the degree of learning and training is also inconsistent, and the diagnostic ability of unknown digital pathological images is still limited. At present, the accuracy of existing algorithm models for diagnosing pathological subtypes of lung adenocarcinoma is not high. Therefore, we need to develop a convolutional neural network model that is highly accurate and suitable for clinical daily work, reducing the impact of repeatable work on doctors' work efficiency

Images