Hosseinpour A, Piranfar A, Harati T, Veyseh S, Soltani M. Clinical applications of generative adversarial networks in medical image to image translation. Int J Radiat Res 2025; 23 (3) :797-807
URL:
http://ijrr.com/article-1-6681-en.html
Department of Mechanical Engineering, K. N. Toosi University of Technology, Tehran, Iran , msoltani@uwaterloo.ca
Abstract: (29 Views)
Generative Adversarial Networks (GANs) have emerged as powerful tools within the realm of deep learning, particularly in the synthesis of artificial images, a capability that holds immense promise in the field of medical image-to-image translation. Recent years have witnessed significant strides in GAN development tailored for cross-domain image translation, largely driven by the availability of extensive datasets containing meticulously annotated medical images. Nonetheless, the process of annotating these images poses a formidable challenge, demanding a substantial number of specialized experts for supervised methods. To surmount this obstacle, cross-modality synthesis techniques have gained traction, offering an efficient approach to mitigate the complexities and costs associated with acquiring paired training data. This paper serves as an introductory exploration into the diverse array of GAN variants employed in image-to-image translation, subsequently delving into their applications within medical imaging. Specifically, it investigates the realms of cross-modality synthesis and conditional image synthesis, shedding light on their potential to revolutionize diagnostic precision and streamline the intricacies of medical imaging processes.
Send email to the article author
