An open source tool with which you can restore faces from old photos is available on GitHub: Generative Facial Prior (GFP) is the name of the tool developed in China, and it is a machine learning model of the Generative Adversarial Network type. (GAN). . According to the project description, GFP-GAN offers “practical algorithms for the realistic reconstruction of faces” in images.
GFP-GAN is open source software licensed under the Apache License 2.0 and comes from a development team at the ARC Lab, where ARC stands for Applied Research Center. Behind it is Chinese social media provider Tencent, which founded the lab in 2019 and claims to be doing media-relevant AI research with it.
Under the hood: multi-stage imaging
the canadian Computer scientist and AI master’s student Louis Bouchard has taken a close look at the underlying technology and features the tool on its “What’s AI” channel on YouTube: while conventional methods for restoring old photos have so far used an AI model that measures the differences between the generated photos and the original, adds the new technology apparently combines information from two complementary AI models and adds photorealistic missing details.
According to Bouchard, the new approach uses a pre-trained version of an AI model, which breaks the image generation process into several phases. With technology, the identity of people can be preserved better than before in photos, among other things, because special attention is paid to facial features, such as the eye and mouth area. The new technology isn’t perfect either, so old or damaged photos in particular are given new details that they didn’t originally contain. Depending on the state of preservation, the reconstructed depictions could look significantly different than the people in the original.
GitHub repository and online demos available
There’s a Colab demo for the project, as well as online demos at Huggingface, Replicate, and BaseTen, among others. If you want to run GFP-GAN yourself, you need Python version 3.7 or higher, PyTorch version 1.7 or higher. The ARC Lab team also recommends installing Anaconda or Miniconda. Optionally, the model can be operated with NVIDIA GPU and CUDA, operation should be possible on both Windows and Linux. The GitHub repository provides installation notes, training tips, and quick inference.
If you want to try the tool, you can load images to optimize them in the browser through a web application called BaseTen or get the Upload the source code from the GitHub repository and integrate the model into your own applications.
Introvert. Beer guru. Communicator. Travel fanatic. Web advocate. Certified alcohol geek. Tv buff. Subtly charming internet aficionado.