Open AccessDomain Translation of Images Using Auto Encoders and Coupled GANs with Unsupervised Learning
Author(s) -
Vigneshkumar Arunachalam,
Harikumar Rajaguru,
Vinubama Krishnamoorthy,
Viswanathan Varadaraj,
R. Pavithra
Publication year - 2021
Publication title -
iop conference series. materials science and engineering
Language(s) - English
Resource type - Journals
eISSN - 1757-899X
pISSN - 1757-8981
DOI - 10.1088/1757-899x/1084/1/012039
Subject(s) - image translation , translation (biology) , computer science , image (mathematics) , domain (mathematical analysis) , nothing , joint (building) , artificial intelligence , encoder , key (lock) , joint probability distribution , pattern recognition (psychology) , computer vision , natural language processing , algorithm , mathematics , mathematical analysis , statistics , architectural engineering , biochemistry , chemistry , philosophy , computer security , epistemology , engineering , gene , operating system , messenger rna
Domain translation of images is nothing but learning from marginal distribution across different to obtain a joint distribution from the listed domains. But there exist a limitless number of joint distributions, but we couldn’t infer that without additional assumptions. It is nothing but changing the entire scenario of the given input image by translating and reconstructing into an entirely new image but with the key features of the input image. Here for the domain translation of images from different domains in video, we are using Coupled-GAN’s for domain translation with making use of shared latent space.