Salvato in:
Dettagli Bibliografici
Autori principali: Hsiao, Chi-Wei, Liu, Yu-Lun, Yang, Cheng-Kun, Kuo, Sheng-Po, Jou, Kevin, Chen, Chia-Ping
Natura: Preprint
Pubblicazione: 2024
Soggetti:
Accesso online:https://arxiv.org/abs/2412.05043
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866917859691593728
author Hsiao, Chi-Wei
Liu, Yu-Lun
Yang, Cheng-Kun
Kuo, Sheng-Po
Jou, Kevin
Chen, Chia-Ping
author_facet Hsiao, Chi-Wei
Liu, Yu-Lun
Yang, Cheng-Kun
Kuo, Sheng-Po
Jou, Kevin
Chen, Chia-Ping
contents While recent works on blind face image restoration have successfully produced impressive high-quality (HQ) images with abundant details from low-quality (LQ) input images, the generated content may not accurately reflect the real appearance of a person. To address this problem, incorporating well-shot personal images as additional reference inputs could be a promising strategy. Inspired by the recent success of the Latent Diffusion Model (LDM), we propose ReF-LDM, an adaptation of LDM designed to generate HQ face images conditioned on one LQ image and multiple HQ reference images. Our model integrates an effective and efficient mechanism, CacheKV, to leverage the reference images during the generation process. Additionally, we design a timestep-scaled identity loss, enabling our LDM-based model to focus on learning the discriminating features of human faces. Lastly, we construct FFHQ-Ref, a dataset consisting of 20,405 high-quality (HQ) face images with corresponding reference images, which can serve as both training and evaluation data for reference-based face restoration models.
format Preprint
id arxiv_https___arxiv_org_abs_2412_05043
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle ReF-LDM: A Latent Diffusion Model for Reference-based Face Image Restoration
Hsiao, Chi-Wei
Liu, Yu-Lun
Yang, Cheng-Kun
Kuo, Sheng-Po
Jou, Kevin
Chen, Chia-Ping
Computer Vision and Pattern Recognition
Machine Learning
Image and Video Processing
While recent works on blind face image restoration have successfully produced impressive high-quality (HQ) images with abundant details from low-quality (LQ) input images, the generated content may not accurately reflect the real appearance of a person. To address this problem, incorporating well-shot personal images as additional reference inputs could be a promising strategy. Inspired by the recent success of the Latent Diffusion Model (LDM), we propose ReF-LDM, an adaptation of LDM designed to generate HQ face images conditioned on one LQ image and multiple HQ reference images. Our model integrates an effective and efficient mechanism, CacheKV, to leverage the reference images during the generation process. Additionally, we design a timestep-scaled identity loss, enabling our LDM-based model to focus on learning the discriminating features of human faces. Lastly, we construct FFHQ-Ref, a dataset consisting of 20,405 high-quality (HQ) face images with corresponding reference images, which can serve as both training and evaluation data for reference-based face restoration models.
title ReF-LDM: A Latent Diffusion Model for Reference-based Face Image Restoration
topic Computer Vision and Pattern Recognition
Machine Learning
Image and Video Processing
url https://arxiv.org/abs/2412.05043