Histopathological Image Segmentation Using U-Net Based Models

Abstract

Medical imaging plays an important role in clinical diagnosis, especially in treatment planning, surgery, and prognosis assessment. It is used to collect potentially life-saving information by looking at human organs without intervention through medical images. Automated segmentation methods can provide reliable diagnostic evidence for specialist physicians in preventive treatment decisions. In this study, different UNet based neural network architectures are investigated for segmentation of histopathological images taken from different organs. The dataset with 19 different organs discussed in the study is segmented using different neural network architectures based on U-Net. As a result of the experiments, the segmentation performances of the architectures are compared and thus a preliminary assessment of real-world problems is carried out.