What is it about?
A simple AI application able to reconstruct the real appearance of an individual starting from some artistic portrait.
Featured Image
Photo by Google DeepMind on Unsplash
Why is it important?
The relevance of the work is in the methodology used to obtain the result. It exploits a modern class of generative models konwn as denoising diffusion models, and an efficient emebedding technique producing the latent representation of a given real image. We first embed the portrait into the latent space, and then use the reverse diffusion model, trained to generate real human faces, to produce the most likely real approximation of the portrait.
Perspectives
The work opens a wide range of fascinating perspectives about the exploration of semantic trajectories in the latent space, the disentanglement of the different aspects of variations, and the possibility of data editing.
Andrea Asperti
Universita degli Studi di Bologna
Read the Original
This page is a summary of: Portrait Reification with Generative Diffusion Models, Applied Sciences, May 2023, MDPI AG,
DOI: 10.3390/app13116487.
You can read the full text:
Contributors
The following have contributed to this page
