Mosaic: Compositional Multi-Concept Erasure via Vector Field Blending
About
Concept erasure has emerged as a key research direction for ensuring safe and ethical image synthesis in Text-to-Image (T2I) models. While existing studies have explored concept erasure across multiple concepts, they typically assume only a single target concept per image, a limitation increasingly exposed by modern flow-based T2I models, which can generate complex scenes with multiple concepts simultaneously. To address this gap, we introduce compositional multi-concept erasure, a new task that aims to simultaneously remove multiple target concepts within a single scene. We propose CoME-Bench, a benchmark for evaluating compositional multi-concept erasure, which covers both intra- and cross-category scenarios. We further propose Mosaic, a novel framework for multi-concept erasure in flow-based T2I models, which exploits the spatial locality of target concepts in the vector field by dynamically constructing concept-specific masks and selectively blending them without additional optimization. Extensive experiments demonstrate that Mosaic effectively removes multiple target concepts in complex compositional scenes while preserving non-target contexts.
Related benchmarks
| Task | Dataset | Result | Rank | |
|---|---|---|---|---|
| Image Compositional Editing | CoME-Bench Cross-Category: Character + Object + Object | ESR0.1366 | 6 | |
| Compositional Image Editing | CoME-Bench Cross-Category: Character + Character + Object | ESR0.2349 | 6 | |
| Compositional Image Editing | CoME-Bench Intra-Category: Character + Character + Character | ESR0.2483 | 6 | |
| Multi-concept Erasure | CoME-Bench Cross-Category: Character + Object | ESR34.19 | 3 | |
| Multi-concept Erasure | CoME-Bench Intra-Category: Character + Character 1.0 (test) | ESR55.23 | 3 | |
| Compositional Image Editing | CoME-Bench Intra-Category: Object + Object + Object | ESR0.0768 | 3 | |
| Object Erasure | CoME-Bench Intra-Category: Object + Object + Object | ESR0.0768 | 3 | |
| Compositional Image Editing | CoME-Bench Intra-Category: Object + Object | ESR0.237 | 3 |