Memorization in 3D Shape Generation: An Empirical Study

TL;DR

This paper introduces an evaluation framework to measure memorization in 3D shape generative models, revealing how data and model design choices influence memorization and proposing strategies to mitigate it.

Contribution

The paper provides the first empirical framework for quantifying memorization in 3D generative models and analyzes how data and model parameters affect memorization levels.

Findings

Memorization depends on data modality and increases with data diversity.

Moderate guidance scale peaks memorization, which can be reduced by longer Vecsets and rotation augmentation.

Strategies to reduce memorization do not degrade generation quality.

Abstract

Generative models are increasingly used in 3D vision to synthesize novel shapes, yet it remains unclear whether their generation relies on memorizing training shapes. Understanding their memorization could help prevent training data leakage and improve the diversity of generated results. In this paper, we design an evaluation framework to quantify memorization in 3D generative models and study the influence of different data and modeling designs on memorization. We first apply our framework to quantify memorization in existing methods. Next, through controlled experiments with a latent vector-set (Vecset) diffusion model, we find that, on the data side, memorization depends on data modality, and increases with data diversity and finer-grained conditioning; on the modeling side, it peaks at a moderate guidance scale and can be mitigated by longer Vecsets and simple rotation augmentation. Together, our framework and analysis provide an empirical understanding of memorization in 3D generative models and suggest simple yet effective strategies to reduce it without degrading generation quality. Our code is available at https://github.com/zlab-princeton/3d_mem.