RAID: Representation Autoencoder for Intrinsic Decomposition
Single-image intrinsic decomposition is an inherently ill-posed problem, requiring the disentanglement of strongly coupled appearance factors from a single RGB observation. Recent methods address this ambiguity with diffusion priors, at the cost of large, iterative, and often stochastic models. We show that accurate decomposition does not require a large generative model. Our key idea is to decouple reconstruction from decomposition on top of a strong pretrained representation. First, an $X \rightarrow X$ teacher learns to reconstruct each intrinsic map from DINOv3 features using a shared decoder, without paired RGB-to-intrinsic supervision. Then, a DINOv3-based student predicts the corresponding teacher representations from RGB through lightweight, modality-specific projections, enabling training on partially annotated datasets and predicting all intrinsics in a single forward pass. \textbf{RAID}, predicts intrinsic properties in the structured latent space, enabling a compact 165M-parameter model, 100x faster than diffusion-based methods. Despite its compact size, it achieves competitive or superior decomposition quality, with improved robustness to input resolution. Additionally, our learned projections yield interpretable, modality-specific feature spaces.
Authors
- Elena Garcés (ORCID: https://orcid.org/0000-0003-3509-8485)
- Giuseppe Vecchio (ORCID: https://orcid.org/0000-0001-5009-4365)
- Tamy Boubekeur (ORCID: https://orcid.org/0000-0001-5985-0921)
- Claudia Cuttano (ORCID: https://orcid.org/0009-0004-9672-507X)
Publication Details
- Journal
- arXiv (Cornell University)
- Published
- 2026-11-18
- DOI
- https://doi.org/10.48550/arxiv.2611.24521
- Primary Topic
- Computer Graphics and Visualization Techniques
- Type
- preprint