Spectral Normalization and Dual Contrastive Regularization for Image-to-Image Translation

arXiv:2304.11319v3 Announce Type: replace
Abstract: Existing image-to-image (I2I) translation methods achieve state-of-the-art performance by incorporating the patch-wise contrastive learning into Generative Adversarial Networks. However, patch-wise contrastive learning only focuses on the local content similarity but neglects the global structure constraint, which affects the quality of the generated images. In this paper, we propose a new unpaired I2I translation framework based on dual contrastive regularization and spectral normalization, namely SN-DCR. To maintain consistency of the global structure and texture, we design …

abstract adversarial art arxiv cs.cv generated generative generative adversarial networks global however image images image-to-image image-to-image translation networks normalization paper performance quality regularization state translation type wise