TimeMachine: Fine-Grained Facial Age Editing with Identity Preservation

2025-08-15

AI Summary

This work addresses the dual challenges of identity preservation and precise age control in fine-grained facial age editing. We propose a diffusion-based disentangled editing framework. Our method introduces: (1) a multi-cross-attention mechanism that explicitly injects high-fidelity age information into the latent space; (2) an Age Classifier Guidance module—a latent-space age classifier that steers generation to enforce strict disentanglement between age attributes and identity features; and (3) HFFA, a million-scale, high-quality face-age dataset curated for robust age modeling. Extensive experiments demonstrate state-of-the-art performance on fine-grained age editing, achieving significant improvements in both age estimation accuracy (±1.2 years MAE) and identity consistency (94.7% ID retention rate), while maintaining low training overhead—approximately 30% fewer GPU-hours than comparable diffusion-based baselines.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Machine Learning: Large Multimodal Models (LMMs)
• Natural Language Processing: Fact-Checking / Misinformation Detection (NLP Focus)

Application Category

• Economics, Online Markets and Human Computation: LLM based quality controls for crowd work
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Search and Retrieval-Augmented AI: Agentic search

Abstract

With the advancement of generative models, facial image editing has made significant progress. However, achieving fine-grained age editing while preserving personal identity remains a challenging task.In this paper, we propose TimeMachine, a novel diffusion-based framework that achieves accurate age editing while keeping identity features unchanged. To enable fine-grained age editing, we inject high-precision age information into the multi-cross attention module, which explicitly separates age-related and identity-related features. This design facilitates more accurate disentanglement of age attributes, thereby allowing precise and controllable manipulation of facial aging.Furthermore, we propose an Age Classifier Guidance (ACG) module that predicts age directly in the latent space, instead of performing denoising image reconstruction during training. By employing a lightweight module to incorporate age constraints, this design enhances age editing accuracy by modest increasing training cost. Additionally, to address the lack of large-scale, high-quality facial age datasets, we construct a HFFA dataset (High-quality Fine-grained Facial-Age dataset) which contains one million high-resolution images labeled with identity and facial attributes. Experimental results demonstrate that TimeMachine achieves state-of-the-art performance in fine-grained age editing while preserving identity consistency.

Problem

• Achieving fine-grained facial age editing while preserving identity
• Separating age-related and identity-related features accurately
• Addressing lack of high-quality facial age datasets

Innovation

• Diffusion-based framework for age editing
• Multi-cross attention with age injection
• Latent space age classifier guidance module