Image Recognition and Diagnosis System of Early Gastric Cancer Based on Artificial Intelligence Algorithm

With the improvement of people's attention to their own health, a large quantity of tumor-related medical data are generated in parallel every day. Artificial intelligence (AI) can summarize and analyze a large quantity of tumor-related data to help doctors work efficiently. Narrow-band imaging (NbI) technology is a brand-new endoscopic imaging technology, which has the function of electronic staining and can highlight the capillary network in the superficial mucosa. Combined with magnifying technology, the fine structure of the superficial mucosa can be displayed more clearly. In this article, an early gastric cancer image recognition algorithm based on convolutional neural network (CNN) is proposed, which provides support for the construction of early gastric cancer image recognition and diagnosis system. Using the collected image data, the quantity of samples is increased through data enhancement on a few data sets, and the precision, recall and error of various models are compared. The results show that the accuracy and recall rate of this algorithm are higher than those of traditional methods, and the recognition error of cancer image in this paper is significantly lower. The model has high accuracy, specificity and sensitivity in the diagnosis of early gastric cancer, and can assist endoscopists in real-time diagnosis during gastroscopy.