What is it about?
Face super-resolution is to convert low-resolution blurred face images into clear and high-resolution face output. At present, the advanced method is to use GAN ways to do face super-resolution. Here, from the earliest GAN model, UR-DGN, to the latest model, FSRCH, we review the development process of taking GAN as a deep model for face super-resolution in the last two years.
Featured Image
Photo by Aleksandr Kozlovskii on Unsplash
Why is it important?
Our findings show that the face super-resolution technology based on GAN is developing towards the direction of preserving the face detailed features of large scale (over 8 times) up-sampling.
Perspectives
The original motivation for this article was an accidental interest: we want to see whether super-resolved face images were helpful for identity identification. This article also promotes us to study more methods of improving image quality through diverse GAN models.
Heng Liu
Read the Original
This page is a summary of: A Survey on GAN-based Face Hallucination with Its Model Development, IET Image Processing, February 2019, the Institution of Engineering and Technology (the IET),
DOI: 10.1049/iet-ipr.2018.6545.
You can read the full text:
Contributors
The following have contributed to this page
