Detection of Lymph Nodes using CNN from Contrast-Enhanced CT Images

Abstract

Detection of lymph nodes in medical practice is important to determine the presence or absence of cancer metastasis or to select the medical operation based on the degree of cancer progression. However, there are problems such that the number of medical doctors is limited or it is difficult to perform the medical diagnosis with high accuracy. This paper tries to solve these problems and proposes a new approach to detect lymph nodes by learning the status of lymph nodes with R2U-Net using divided patches of contrast-enhanced CT images. It is shown that the detection of lymph node becomes better by introducing the integration processing.