Cell Image Segmentation by Integrating Generative Adversarial Network for Each Class

Abstract

Human experts segment cell images manually now, and the criterion for segmentation varies on each expert. As a result, subjective results are obtained. If we develop an automatic segmentation method, we can obtain objective results by the same criteria. This paper proposes a cell image segmentation method using Generative Adversarial Network (GAN) with multiple different roles. The proposed method improved the segmentation accuracy in comparison to conventional pix2pix.