Abstract
Abstract: Pokémon is a video game series wherein players capture and train fauna that are known as Pokémons. These creatures vary in colour, shape, size, and have distinct personalities and skillsets. After the Pokémons are trained by their coaches (who are the players), they are used to fight other Pokémons so that the winning player could achieve the Pokémon Champion status. Since its debut in 1996, Pokémon has become the highest grossing media franchise with over $100 billion in revenue through the sale of books and merchandise along with television programming and movies. This paper discusses the implementation of the Wasserstein Generative Adversarial Network (WGAN) to create artificial bipedal Pokémon images. This is achieved by training a WGAN on a custom-built training dataset of Generation I bipedal Pokémon pictures. The effectiveness of the WGAN for outputting artificial Pokémon images is evaluated by using the loss curves and by visually inspecting the outputted images.
Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
