AgeBooth: Controllable Facial Aging and Rejuvenation via Diffusion Models

2025-10-07

AI Summary

Existing cross-age face editing methods face two key challenges: (1) difficulty in precisely controlling target age while preserving identity consistency, and (2) heavy reliance on large-scale paired cross-age image datasets for fine-tuning, entailing high data acquisition costs. To address these, we propose a pair-free, age-controllable diffusion-based editing framework. First, we introduce an age-specific fine-tuning strategy that embeds discrete age labels as conditional prompts and applies dynamic blending. Second, we design SVDMix—a matrix-based fusion mechanism that jointly optimizes LoRA adapter parameters and an identity-personalized diffusion model. Extensive experiments demonstrate that our method significantly outperforms state-of-the-art approaches in age accuracy (MAE reduced by 2.1 years), identity fidelity (ID similarity improved by 18.7%), and perceptual realism. Moreover, it enables high-fidelity aging and rejuvenation editing at arbitrary intermediate ages without requiring paired training data.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Search and Optimization: Mixed Discrete/Continuous Search
• Machine Learning: Large Multimodal Models (LMMs)

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Economics, Online Markets and Human Computation: LLM based quality controls for crowd work

Abstract

Recent diffusion model research focuses on generating identity-consistent images from a reference photo, but they struggle to accurately control age while preserving identity, and fine-tuning such models often requires costly paired images across ages. In this paper, we propose AgeBooth, a novel age-specific finetuning approach that can effectively enhance the age control capability of adapterbased identity personalization models without the need for expensive age-varied datasets. To reduce dependence on a large amount of age-labeled data, we exploit the linear nature of aging by introducing age-conditioned prompt blending and an age-specific LoRA fusion strategy that leverages SVDMix, a matrix fusion technique. These techniques enable high-quality generation of intermediate-age portraits. Our AgeBooth produces realistic and identity-consistent face images across different ages from a single reference image. Experiments show that AgeBooth achieves superior age control and visual quality compared to previous state-of-the-art editing-based methods.

Problem

• Enhancing age control in identity-preserving facial generation
• Reducing dependency on costly age-varied paired datasets
• Generating realistic intermediate-age portraits from single reference

Innovation

• Age-specific finetuning for adapter-based identity models
• Age-conditioned prompt blending for linear aging control
• SVDMix LoRA fusion strategy for intermediate-age generation