Rectifying Latent Space for Generative Single-Image Reflection Removal
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2025
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866909947245101056 |
|---|---|
| author | Li, Mingjia Hu, Jin Wang, Hainuo Hu, Qiming Wang, Jiarui Guo, Xiaojie |
| author_facet | Li, Mingjia Hu, Jin Wang, Hainuo Hu, Qiming Wang, Jiarui Guo, Xiaojie |
| contents | Single-image reflection removal is a highly ill-posed problem, where existing methods struggle to reason about the composition of corrupted regions, causing them to fail at recovery and generalization in the wild. This work reframes an editing-purpose latent diffusion model to effectively perceive and process highly ambiguous, layered image inputs, yielding high-quality outputs. We argue that the challenge of this conversion stems from a critical yet overlooked issue, i.e., the latent space of semantic encoders lacks the inherent structure to interpret a composite image as a linear superposition of its constituent layers. Our approach is built on three synergistic components, including a reflection-equivariant VAE that aligns the latent space with the linear physics of reflection formation, a learnable task-specific text embedding for precise guidance that bypasses ambiguous language, and a depth-guided early-branching sampling strategy to harness generative stochasticity for promising results. Extensive experiments reveal that our model achieves new SOTA performance on multiple benchmarks and generalizes well to challenging real-world cases. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2512_06358 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Rectifying Latent Space for Generative Single-Image Reflection Removal Li, Mingjia Hu, Jin Wang, Hainuo Hu, Qiming Wang, Jiarui Guo, Xiaojie Computer Vision and Pattern Recognition Single-image reflection removal is a highly ill-posed problem, where existing methods struggle to reason about the composition of corrupted regions, causing them to fail at recovery and generalization in the wild. This work reframes an editing-purpose latent diffusion model to effectively perceive and process highly ambiguous, layered image inputs, yielding high-quality outputs. We argue that the challenge of this conversion stems from a critical yet overlooked issue, i.e., the latent space of semantic encoders lacks the inherent structure to interpret a composite image as a linear superposition of its constituent layers. Our approach is built on three synergistic components, including a reflection-equivariant VAE that aligns the latent space with the linear physics of reflection formation, a learnable task-specific text embedding for precise guidance that bypasses ambiguous language, and a depth-guided early-branching sampling strategy to harness generative stochasticity for promising results. Extensive experiments reveal that our model achieves new SOTA performance on multiple benchmarks and generalizes well to challenging real-world cases. |
| title | Rectifying Latent Space for Generative Single-Image Reflection Removal |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2512.06358 |