Pluralistic Aging Diffusion Autoencoder

TL;DR

This paper introduces PADA, a novel diffusion autoencoder that generates diverse, high-quality face aging results by combining diffusion models with probabilistic aging embeddings guided by CLIP, addressing the ill-posed nature of face aging.

Contribution

It proposes a CLIP-driven diffusion autoencoder with probabilistic aging embeddings to produce diverse aging patterns conditioned on open-world texts and unseen images.

Findings

Generates diverse plausible aging results

Achieves high-quality face aging conditioned on open-world texts

Outperforms existing methods in diversity and quality

Abstract

Face aging is an ill-posed problem because multiple plausible aging patterns may correspond to a given input. Most existing methods often produce one deterministic estimation. This paper proposes a novel CLIP-driven Pluralistic Aging Diffusion Autoencoder (PADA) to enhance the diversity of aging patterns. First, we employ diffusion models to generate diverse low-level aging details via a sequential denoising reverse process. Second, we present Probabilistic Aging Embedding (PAE) to capture diverse high-level aging patterns, which represents age information as probabilistic distributions in the common CLIP latent space. A text-guided KL-divergence loss is designed to guide this learning. Our method can achieve pluralistic face aging conditioned on open-world aging texts and arbitrary unseen face images. Qualitative and quantitative experiments demonstrate that our method can generate more diverse and high-quality plausible aging results.