Cancer is termed as one of the deadliest disease, and it is becoming a major health problem in the world. This deadly disease can be cured, if it is found at earlier stages. Medical imaging plays an important role in finding this type of diseases and helps in treatment planning. Automated lesion/tumor segmentation is an important and challenging clinical diagnostic task because of tumor’s different shape, volume, contrast, and locations. Since deep learning is promising in many applications, thus motivating us to apply in this important task. In this paper, we propose to do automated tumor segmentation with two challenge dataset named as BraTs 2017 brain tumor and ISIC 2018 skin lesion by using Segmentor Adversarial Network (SegAN), inspired from classical GAN.
S. Saj T. K, Vishvanathan, S., and Dr. Soman K. P., “Performance Analysis of Segmentor Adversarial Network (SegAN) on Bio-Medical Images for Image Segmentation”, in Advances in Automation, Signal Processing, Instrumentation, and Control, 2021, pp. 751-758.