Frequency Domain Image Translation: More Photo-realistic, Better Identity-preserving
Image-to-image translation has been revolutionized with GAN-based methods. However, existing methods lack the ability to preserve the identity of the source domain. As a result, synthesized images can often over-adapt to the reference domain, losing important structural characteristics and suffering from suboptimal visual quality. To solve these challenges, we propose a novel frequency domain image translation (FDIT) framework, exploiting frequency information for enhancing the image generation process. Our key idea is to decompose the image into low-frequency and high-frequency components, where the high-frequency feature captures object structure akin to the identity. Our training objective facilitates the preservation of frequency information in both pixel space and Fourier spectral space. We broadly evaluate FDIT across five large-scale datasets and multiple tasks including image translation and GAN inversion. Extensive experiments and ablations show that FDIT effectively preserves the identity of the source image, and produces photo-realistic images. FDIT establishes state-of-the-art performance, reducing the average FID score by 5.6% compared to the previous best method.
Code (1)
Tasks
Image GenerationImage-to-Image TranslationTranslationSimilar Papers 제목 키워드 기반
Spectrum Translation for Refinement of Image Generation (STIG) Based on Contrastive Learning and Spectral Filter Profile
Currently, image generation and synthesis have remarkably progressed with generative models. Despite photo-realistic results, intrinsic discrepancies are still observed in the frequency domain. The spectral discrepancy a…
Contrastive LearningFace SwappingImage GenerationImage-to-Image Translation+1Learning Latent Representations for Image Translation using Frequency Distributed CycleGAN
This paper presents Fd-CycleGAN, an image-to-image (I2I) translation framework that enhances latent representation learning to approximate real data distributions. Building upon the foundation of CycleGAN, our approach i…
Representation LearningStyle TransferHigh-Resolution Photorealistic Image Translation in Real-Time: A Laplacian Pyramid Translation Network
Existing image-to-image translation (I2IT) methods are either constrained to low-resolution images or long inference time due to their heavy computational burden on the convolution of high-resolution feature maps. In thi…
4kAttributeColor ManipulationGPU+4Wavelet-based Unsupervised Label-to-Image Translation
Semantic Image Synthesis (SIS) is a subclass of image-to-image translation where a semantic layout is used to generate a photorealistic image. State-of-the-art conditional Generative Adversarial Networks (GANs) need a hu…
Image GenerationImage-to-Image TranslationMultimodal Unsupervised Image-To-Image TranslationTranslation+1Old Photo Restoration via Deep Latent Space Translation
We propose to restore old photos that suffer from severe degradation through a deep learning approach. Unlike conventional restoration tasks that can be solved through supervised learning, the degradation in real photos …
Image RestorationTranslationTriplet