NVIDIA researchers, led by Ting-Chun Wang, have developed a new deep learning-based, artificial intelligence system that can generate photo-realistic images using only high-level labels, and at the same time creating a virtual environment that allows the user to modify a scene interactively. The method, which uses a conditional generative adversarial network (GAN), has the potential to revolutionize how visual algorithms, including those for medical imaging, are trained. “Conditional GANs have enabled a variety of applications, but the results are often limited to low-resolution and still far from realistic. In this work, we generate 2048×1024 visually appealing results with a novel adversarial loss, as well as new multi-scale generator and discriminator architectures,” said the researchers in their paper. Click here for the first image in today’s viral picture gallery. Continue reading for the five most popular viral videos today, including one where “As Seen on TV” pest control gadgets are tested.
