WaveFace: Authentic Face Restoration with Efficient Frequency Recovery

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Miao, Yunqi, Deng, Jiankang, Han, Jungong
Format: Preprint
Veröffentlicht: 2024
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866929282249392128
author Miao, Yunqi
Deng, Jiankang
Han, Jungong
author_facet Miao, Yunqi
Deng, Jiankang
Han, Jungong
contents Although diffusion models are rising as a powerful solution for blind face restoration, they are criticized for two problems: 1) slow training and inference speed, and 2) failure in preserving identity and recovering fine-grained facial details. In this work, we propose WaveFace to solve the problems in the frequency domain, where low- and high-frequency components decomposed by wavelet transformation are considered individually to maximize authenticity as well as efficiency. The diffusion model is applied to recover the low-frequency component only, which presents general information of the original image but 1/16 in size. To preserve the original identity, the generation is conditioned on the low-frequency component of low-quality images at each denoising step. Meanwhile, high-frequency components at multiple decomposition levels are handled by a unified network, which recovers complex facial details in a single step. Evaluations on four benchmark datasets show that: 1) WaveFace outperforms state-of-the-art methods in authenticity, especially in terms of identity preservation, and 2) authentic images are restored with the efficiency 10x faster than existing diffusion model-based BFR methods.
format Preprint
id arxiv_https___arxiv_org_abs_2403_12760
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle WaveFace: Authentic Face Restoration with Efficient Frequency Recovery
Miao, Yunqi
Deng, Jiankang
Han, Jungong
Computer Vision and Pattern Recognition
Although diffusion models are rising as a powerful solution for blind face restoration, they are criticized for two problems: 1) slow training and inference speed, and 2) failure in preserving identity and recovering fine-grained facial details. In this work, we propose WaveFace to solve the problems in the frequency domain, where low- and high-frequency components decomposed by wavelet transformation are considered individually to maximize authenticity as well as efficiency. The diffusion model is applied to recover the low-frequency component only, which presents general information of the original image but 1/16 in size. To preserve the original identity, the generation is conditioned on the low-frequency component of low-quality images at each denoising step. Meanwhile, high-frequency components at multiple decomposition levels are handled by a unified network, which recovers complex facial details in a single step. Evaluations on four benchmark datasets show that: 1) WaveFace outperforms state-of-the-art methods in authenticity, especially in terms of identity preservation, and 2) authentic images are restored with the efficiency 10x faster than existing diffusion model-based BFR methods.
title WaveFace: Authentic Face Restoration with Efficient Frequency Recovery
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2403.12760