Identity Preserving Latent Diffusion for Brain Aging Modeling

2025-03-11

AI Summary

To address the challenge of preserving subject identity consistency in cross-age brain image synthesis, this paper proposes the Identity-Preserving Longitudinal Diffusion Model (IP-LDM). Built upon a latent-space diffusion framework, IP-LDM integrates an identity encoder with age-conditional modulation and introduces, for the first time, triplet contrastive regularization to explicitly constrain latent representations for identity fidelity—thereby overcoming the limitations of conventional i.i.d. assumptions in modeling temporal consistency. Evaluated on multi-stage brain MRI data spanning infancy to old age, IP-LDM generates anatomically plausible and photorealistic images, achieving a 23.6% improvement in subject identity similarity and a 19.4% gain in cross-timepoint reconstruction consistency over state-of-the-art conditional generative models. This work establishes a novel, interpretable, and identity-robust paradigm for longitudinal medical image modeling.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Machine Learning: Large Multimodal Models (LMMs)
• Natural Language Processing: Generation

Application Category

• User Modeling, Personalization and Recommendation: Studies of user behavior, including longitudinal effects of personalized systems
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Social Networks and Social Media: Generative AI / large language models and their impact on social systems

Abstract

Structural and appearance changes in brain imaging over time are crucial indicators of neurodevelopment and neurodegeneration. The rapid advancement of large-scale generative models provides a promising backbone for modeling these complex global and local changes in brain images, such as transforming the age of a source image to a target age. However, current generative models, typically trained on independently and identically distributed (i.i.d.) data, may struggle to maintain intra-subject spatiotemporal consistency during transformations. We propose the Identity-Preserving Longitudinal Diffusion Model (IP-LDM), designed to accurately transform brain ages while preserving subject identity. Our approach involves first extracting the identity representation from the source image. Then, conditioned on the target age, the latent diffusion model learns to generate the age-transformed target image. To ensure consistency within the same subject over time, we regularize the identity representation using a triplet contrastive formulation. Our experiments on both elderly and infant brain datasets demonstrate that our model outperforms existing conditional generative models, producing realistic age transformations while preserving intra-subject identity.

Problem

• Model brain aging while preserving subject identity
• Maintain intra-subject spatiotemporal consistency in transformations
• Generate realistic age transformations using latent diffusion

Innovation

• Identity-Preserving Longitudinal Diffusion Model (IP-LDM)
• Extracts identity representation from source image
• Regularizes identity with triplet contrastive formulation