What's in a Decade? Transforming Faces Through Time
This work addresses the challenge of decade-specific portrait synthesis for applications in historical analysis and creative media, though it is incremental as it builds on existing image translation techniques.
The authors tackled the problem of visually characterizing people across decades by assembling the Faces Through Time dataset and developing a framework to resynthesize portrait images across time, showing that their method outperforms state-of-the-art image-to-image translation and editing models.
How can one visually characterize people in a decade? In this work, we assemble the Faces Through Time dataset, which contains over a thousand portrait images from each decade, spanning the 1880s to the present day. Using our new dataset, we present a framework for resynthesizing portrait images across time, imagining how a portrait taken during a particular decade might have looked like, had it been taken in other decades. Our framework optimizes a family of per-decade generators that reveal subtle changes that differentiate decade--such as different hairstyles or makeup--while maintaining the identity of the input portrait. Experiments show that our method is more effective in resynthesizing portraits across time compared to state-of-the-art image-to-image translation methods, as well as attribute-based and language-guided portrait editing models. Our code and data will be available at https://facesthroughtime.github.io