[1Z3] [General Session] 2. Machine Learning

[1Z3-01] A Proposal of a loss function of GAN to generate various images

〇Ryoji Kodama¹, Tsuyoshi Nakamura¹, Masayoshi Kanoh², Koji Yamada³ (1. Nagoya Institute of Technology, 2. Chukyo University, 3. Institute of Advanced Media Arts and Sciences)

Keywords:Generative Model, DCGAN, illustration, loss function

We propose a new loss function based on variance of generated images that we introduce to GAN(Generative
Adversarial Networks). Recent image-generation methods adopt a neural-network based generative model. Auto-
generators of illustrations can contribute to assist creative activities and entertainment. This paper focuses on
supression of collapse and its benefit to GAN training. Using our new technique, we attempted to generate various
images of illustrations.