Demystifying Foreground-Background Memorization in Diffusion Models
This addresses the issue of memorization in diffusion models for AI safety and privacy, but it is incremental as it builds on existing detection methods with a new metric.
The paper tackled the problem of memorization in diffusion models by proposing FB-Mem, a segmentation-based metric that quantifies partial memorization in small image regions and reveals complex patterns beyond one-to-one prompt-image pairs, showing memorization is more pervasive than previously thought and that current mitigation methods fail to eliminate local memorization.
Diffusion models (DMs) memorize training images and can reproduce near-duplicates during generation. Current detection methods identify verbatim memorization but fail to capture two critical aspects: quantifying partial memorization occurring in small image regions, and memorization patterns beyond specific prompt-image pairs. To address these limitations, we propose Foreground Background Memorization (FB-Mem), a novel segmentation-based metric that classifies and quantifies memorized regions within generated images. Our method reveals that memorization is more pervasive than previously understood: (1) individual generations from single prompts may be linked to clusters of similar training images, revealing complex memorization patterns that extend beyond one-to-one correspondences; and (2) existing model-level mitigation methods, such as neuron deactivation and pruning, fail to eliminate local memorization, which persists particularly in foreground regions. Our work establishes an effective framework for measuring memorization in diffusion models, demonstrates the inadequacy of current mitigation approaches, and proposes a stronger mitigation method using a clustering approach.