Optimization Research and E-commerce Design of Gastrointestinal Electron Microscopic Endoscopy Images Based on Convolutional Neural Networks

Abstract

The research team backtracked the gastrointestinal electron microscopy manifestations of patients based on the pathological classification of clinical institutions. Through data backtracking, we obtained the diagnostic significance images of the patients. The diagnostic significance images were classified and implemented using convolutional neural networks, with Class A being benign lesions, Class C being malignant lesions, and Class B being non lesions. Based on existing data, we obtained a confusion matrix, and a report was made on the parameters of convolutional neural networks. This convolutional neural network has decent accuracy and can preliminarily meet the needs of clinical institutions, providing a powerful tool for clinical work.