Searching for esophageal cancer lesion area recognition modeling method based on evolutionary neural network structure

Abstract

The invention discloses a method for identifying and modeling esophageal cancer lesion regions based on evolutionary neural network structure search, which relates to the technical field of image pattern recognition and medical images, and includes the following steps: S1: collecting and marking the esophagus for training neural network models Image dataset; S2: Construct a neural network structure search space for esophageal cancer lesion area recognition; S3: Train a super network model for esophageal cancer lesion area recognition; S4: Use evolutionary algorithm to search for the optimal super network model Neural network structure; S5: Fine-tune the searched neural network structure and predict the lesion area on the newly input esophageal image. The present invention eliminates the dependence of neural network structure design on expert experience in the intelligent identification task of esophageal cancer, making the deep neural network method easier to use in the area identification of esophageal cancer lesions.

Description

technical field [0001] The invention relates to the technical field of image pattern recognition and medical images, in particular to a method for identifying and modeling esophageal cancer lesion regions based on evolutionary neural network structure search. Background technique [0002] Esophageal cancer (Esophageal Cancer, EC) is a malignant disease with a high mortality rate [A. K. Rustgiand H. B. El-Serag, “Esophageal cancer,” New England Journal of Medicine, vol. 371, no. 26, pp. 2499–2509 , 2014.], studies have shown that its 5-year survival rate is less than 20% [ X.-X.Chen, Q. Zhong, Y. Liu, S.-M. Yan, Z.-H. Chen, S.-Z . Jin, T.-L. Xia, R.-Y.Li, A.-J. Zhou, Z. Su et al., “Genomic comparison of esophageal squamous cellcarcinoma and its precursor lesions by multi-region whole-exome sequencing ,” Nature Communications, vol. 8, no. 1, pp. 1–12, 2017.]. Gastroscopy has been widely used in the diagnosis of early esophageal cancer, which can provide guidance for early in...

AI Technical Summary

Problems solved by technology

[0006] In general: the convolutional neural network method has been widely used in the identification of esophageal cancer lesion regions, but the construction of the network model has a strong dependence on the experience of doctors and neural network experts

Images