Abstract

Info

Author(s):: GONG Jinchang; ZHAO Shangyi; WANG Yuanjun; Institute of Medical Imaging Engineering, University of Shanghai for Science and Technology, Shanghai 200093, China

Keywords:: Keywords: medical image segmentation; deep learning; convolutional neural network; review

Abstract:: Abstract: Medical image segmentation is a key step in the quantitative analysis of medical images. Therefore, segmentation of lesions is of great significance for clinical diagnosis. Based on the problems of traditional segmentation methods such as excessive dependence on the prior knowledge of medical science and errors in subjective assessment, the researchers propose a deep learning-based method for lesion segmentation. Herein the research progress of convolutional neural network algorithm in lesion segmentation is summarized. First, the basic structure and architecture of convolutional neural network are expounded. Secondly, the application of deep learning in lesion segmentation, including the detection and classification of pulmonary nodules and the segmentation of brain tumors and breast lesions, are introduced. Finally, the advantages and disadvantages existed in this research are analyzed and the development direction of deep learning is prospected.