Brain tumor image generations using Deep Convolutional Generative adversarial networks

(DCGAN)

Abstract

Brain tumors are one of the leading causes of cancer related death around the world. Multiple time points of the single or multimodal medical images of the same patient are required to study tumor development prediction. Owing to the generation of the large-scale data and development in the field of deep learning, amazing results have been achieved in the field of computer vision especially in the domain area of medical imaging. In this field of medical imaging the generation of practical data is the most challenging and laborious activity. To overcome this, we have used the Deep convolutional generative adversarial networks (DCGAN) machine learning algorithm to generate the fake samples. In this research paper, we extracted the salient features of DCGAN and practically demonstrated and implemented it to generate the fake scans of the brain tumors and found it to outperform other types of generative adversarial networks (GANs). We believe that the generation of this fake data while presented to expert physicians will be considered as realistic data.