Faces of a person across years, generated through personalized re-aging like a time machine

What is it about?

Inspired by The Irishman, we present MyTimeMachine, a generative AI system that adapts a pre-trained StyleGAN aging encoder using only ~50 personal selfies. It produces personalized, identity-preserving, and temporally consistent age regression and progression in both images and videos, without the need for costly equipment or manual modeling.

Featured Image

Photo by A Chosen Soul on Unsplash

Photo by A Chosen Soul on Unsplash

Why is it important?

Previous approaches either (1) rely on costly human labor and professional camera setups to explicitly model faces for re-aging (e.g., via meshes), or (2) fail to personalize the re-aging process, producing plausible but generic—and often inaccurate—results. By contrast, our method, MyTimeMachine (MyTM), employs generative AI to personalize facial age transformation using only ~50 in-the-wild selfies. Built atop a pre-trained global aging encoder and augmented via a novel Adapter Network, MyTM achieves both age regression (de-aging) and age progression (aging) in images and videos—delivering high-fidelity, identity-preserving, and temporally consistent results without the need for specialized equipment or manual modeling.