Computer Aided Annotation of Early Esophageal Cancer in Gastroscopic Images based on Deeplabv3+ Network

摘要: The diagnoses of Early Esophageal Cancer (EEC) based on gastroscopic images is a challenging task in clinic, which relies heavily subjective artificial detection and annotation. As result, computer aided diagnosis (CAD) methods that support the clinicians become highly attractive. In this paper, we proposed CAD method realized automatic annotation EEC lesions images. initially utilized an advanced Deep Learning (DL) network Deeplabv3+ to obtain preliminary prediction regions. Then, post-processing step referenced clinical requirements was designed applied get final results. Totally 3190 732 patients were used work. experimental results show rate our 97.07%, mean Dice Similarity Coefficient (DSC) 74.01%, are higher than those other state-of-the-are DL-based methods. addition, false positive output fewer. Therefore, offers good potential aid EEC.