Summary of Ref-ldm: a Latent Diffusion Model For Reference-based Face Image Restoration, by Chi-wei Hsiao et al.
ReF-LDM: A Latent Diffusion Model for Reference-based Face Image Restoration
by Chi-Wei Hsiao, Yu-Lun Liu, Cheng-Kun Yang, Sheng-Po Kuo, Kevin Jou, Chia-Ping Chen
First submitted to arxiv on: 6 Dec 2024
Categories
- Main: Computer Vision and Pattern Recognition (cs.CV)
- Secondary: Machine Learning (cs.LG); Image and Video Processing (eess.IV)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary The proposed ReF-LDM model adapts the Latent Diffusion Model to generate high-quality face images conditioned on a low-quality input image and multiple high-quality reference images. The integration of CacheKV and timestep-scaled identity loss enables the model to focus on learning facial features. The FFHQ-Ref dataset, comprising 20,405 HQ face images with corresponding reference images, serves as both training and evaluation data for reference-based face restoration models. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary A new approach in blind face image restoration aims to generate high-quality images that accurately reflect a person’s real appearance. To achieve this, the ReF-LDM model uses multiple well-shot personal images as references, alongside a low-quality input image. This innovative strategy can improve the accuracy of restored images by incorporating facial features from multiple sources. |
Keywords
* Artificial intelligence * Diffusion model
