Scenimefy: Learning to Craft Anime Scene via Semi-Supervised Image-to-Image Translation

TL;DR

Scenimefy is a semi-supervised framework that transforms real-world images into anime scenes, leveraging pseudo data, segmentation, and contrastive loss to improve stylization, details, and semantic consistency.

Contribution

It introduces a novel semi-supervised image translation method using structure-consistent pseudo data and a high-resolution anime dataset, advancing anime scene rendering quality.

Findings

Outperforms state-of-the-art methods in perceptual quality

Achieves better semantic preservation and stylization

Provides a new high-resolution anime scene dataset

Abstract

Automatic high-quality rendering of anime scenes from complex real-world images is of significant practical value. The challenges of this task lie in the complexity of the scenes, the unique features of anime style, and the lack of high-quality datasets to bridge the domain gap. Despite promising attempts, previous efforts are still incompetent in achieving satisfactory results with consistent semantic preservation, evident stylization, and fine details. In this study, we propose Scenimefy, a novel semi-supervised image-to-image translation framework that addresses these challenges. Our approach guides the learning with structure-consistent pseudo paired data, simplifying the pure unsupervised setting. The pseudo data are derived uniquely from a semantic-constrained StyleGAN leveraging rich model priors like CLIP. We further apply segmentation-guided data selection to obtain high-quality pseudo supervision. A patch-wise contrastive style loss is introduced to improve stylization and fine details. Besides, we contribute a high-resolution anime scene dataset to facilitate future research. Our extensive experiments demonstrate the superiority of our method over state-of-the-art baselines in terms of both perceptual quality and quantitative performance.