NFT artwork generation using oscillatory activation functions in GANs

Abstract

The concept of digital ownership is not new, and has been widely used in gaming contexts to allow players to customize their experiences via profile pictures, skins, upgrades and add-ons. In this paper we propose a novel model for NFT generation which uses Oscillatory activation function instead of other mainstream activation functions. Here, we are using a combination of GCU and ReLU to train the model and the subsequently use for the prediction. We have used the Bored Apes Yacht Club Dataset available here. This dataset contains 10,000 images of famous NFTs from Bored Apes Yacht Club. NFTs will accelerate the growth of the cryptocurrency space outside of finance, and will bring novel ideas and approaches from new sets of creators, artists, collectors of digital items, developers and more.