Knowledge-driven non-small cell lung cancer subtype classification method and system

By adopting a knowledge-driven deep learning model in medical image processing, combining CBAM module and decoding module to extract and reconstruct three-dimensional image features, the problems of feature extraction limitations and manual labeling dependence in the prior art are solved, and a highly accurate classification of non-small cell lung cancer subtypes is achieved.

CN119741542APending Publication Date: 2025-04-01ARMY MEDICAL UNIV +1
0 Cites 0 Cited by

Patent Information

Application Number
CN202411808898.5
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-12-10
Publication Date
2025-04-01

AI Technical Summary

Technical Problem

In the prior art, feature extraction depends on two-dimensional slices in medical image processing. Deep learning models have limitations in using three-dimensional image feature extraction. At the same time, subtype classification depends on doctors' manual annotation, resulting in limited number of data objects.

Method used

A knowledge-driven non-small cell lung cancer subtype classification method is proposed, using a deep learning model of non-small cell lung cancer subtype classification branch and tumor mask reconstruction branch, and extract and reconstruct three-dimensional image features through the CBAM module and the decoding module, generate masks and perform subtype classification.

Benefits of technology

It effectively realizes the classification of non-small cell lung cancer subtypes, reduces dependence on experienced doctors, improves classification accuracy, and provides auxiliary diagnostic tools for medical institutions with limited resources.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

The invention belongs to the technical field of image processing, and particularly relates to a non-small cell lung cancer subtype classification method and system based on knowledge driving. Comprising the following steps: acquiring a to-be-predicted image, preprocessing the to-be-predicted image, and inputting the preprocessed image into a trained non-small cell lung cancer subtype classification model for processing to obtain an image subtype classification result; the model comprises a non-small cell lung cancer subtype classification branch and a non-small cell lung cancer tumor mask reconstruction branch; the subtype classification branch is a 3D ResNet34 model in which a CBAM module is added to a coding path; the tumor mask reconstruction branch and the subtype classification branch share a coding path, and four decoding modules and sigmoid functions are added behind the coding path of the mask reconstruction branch; the classification accuracy is high, the suspected subtype classification result obtained through the subtype classification method can serve as an effective judgment basis, doctors are assisted in better lung cancer subtype diagnosis, and good application prospects are achieved.
Need to check novelty before this filing date? Find Prior Art