Answers are grounded in this paper and cite their sources.
DOTA2 is a popular game in the E-sport game industry with a high market value.Developing a variety of in-game character traits can increase the company's revenue and player satisfaction.Many studies can demonstrate the effectiveness of Generative Adversarial Networks (GAN) applications to create cartoon images or recreate portraits.Thus, this research aims to study the effectiveness of the StyleGAN application to be used in recreating and redesigning DOTA2 characters.To create new characters using a small number of DOTA2 original character images, we use a few-shot Image Generation via Cross-domain Correspondence.The dataset used in training includes 111 thumbnail images of DOTA2 hero characters.Three models were trained and experimented with.Manual observation showed that models adapted from the FFHQ model correctly adapted a person's eyes and nose to the character's eyes and nose; however, the mouth part was erased.The other models could only create half the image but still correspond to the character traits.The images created by the three models were quite dark and fresh, which also looked like the images of a DOTA2 hero.As for the qualitative evaluation, it was found that users preferred the FFHQ model the most, followed by the Church model and the Horses model, respectively.The main consideration factors were how much the generated images are similar in style to DOTA2.
No takes yet. Share an insight, caveat, or question.
Angkawinijwong et al. (2023) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: