Research on ME- NBI Gastric Lesion Recognition System based on Improved UNet Structure

Early gastric cancer develops through inflammation, intestinalization, low-grade neoplasia and other developmental lesions, yet most patients are already at an advanced stage of gastric cancer when detected. In order to effectively identify lesion sites from endoscopic images acquired by magnified endoscopy combined with narrow-band imaging (ME-NBI), this paper introduces deep learning into an intelligent assisted lesion diagnosis system for endoscopic images, using an improved UNet-based network for accurate lesion site segmentation. The experimental results show that the reconciliation parameter F-score of the accuracy and recall of the proposed model in this paper can reach 96%, indicating that the improved method has a better ability to identify early cancer and precancerous lesions in the stomach.