# Intrinsic Image Decomposition using Paradigms

arXiv: https://arxiv.org/abs/2011.10512 · HTML: https://mlsanity.com/paper/2011.10512/

## Key facts

- Cited by 17 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 20 Nov 2020.

## Built on by (most cited first, 17 of 17)

- [Intrinsic Image Diffusion for Indoor Single-view Material Estimation](https://mlsanity.com/paper/2312.12274/) (2023)
- [Generative Models: What Do They Know? Do They Know Things? Let's Find Out!](https://mlsanity.com/paper/2311.17137/) (2023)
- [StyleGAN knows Normal, Depth, Albedo, and More](https://mlsanity.com/paper/2306.00987/) (2023)
- [LumiNet: Latent Intrinsics Meets Diffusion Models for Indoor Scene Relighting](https://mlsanity.com/paper/2412.00177/) (2024)
- [UrbanIR: Large-Scale Urban Scene Inverse Rendering from a Single Video](https://mlsanity.com/paper/2306.09349/) (2023)
- [Latent Intrinsics Emerge from Training to Relight](https://mlsanity.com/paper/2405.21074/) (2024)
- [Measured Albedo in the Wild: Filling the Gap in Intrinsics Evaluation](https://mlsanity.com/paper/2306.15662/) (2023)
- [Boosting 3D Object Generation through PBR Materials](https://mlsanity.com/paper/2411.16080/) (2024)
- [StyLitGAN: Prompting StyleGAN to Produce New Illumination Conditions](https://mlsanity.com/paper/2205.10351/) (2022)
- [Creating a Forensic Database of Shoeprints from Online Shoe Tread Photos](https://mlsanity.com/paper/2205.02361/) (2022)
- [DPF: Learning Dense Prediction Fields with Weak Supervision](https://mlsanity.com/paper/2303.16890/) (2023)
- [Long Scale Error Control in Low Light Image and Video Enhancement Using Equivariance](https://mlsanity.com/paper/2206.01334/) (2022)
- [SIRfyN: Single Image Relighting from your Neighbors](https://mlsanity.com/paper/2112.04497/) (2021)
- [SAIL: Self-supervised Albedo Estimation from Real Images with a Latent Diffusion Model](https://mlsanity.com/paper/2505.19751/) (2025)
- [Shape from Shading for Robotic Manipulation](https://mlsanity.com/paper/2304.11824/) (2023)
- [Intrinsic Image Decomposition Using Point Cloud Representation](https://mlsanity.com/paper/2307.10924/) (2023)
- [Texture-aware Intrinsic Image Decomposition with Model- and Learning-based Priors](https://mlsanity.com/paper/2509.09352/) (2025)

## Abstract (arXiv)

Intrinsic image decomposition is the classical task of mapping image to albedo. The WHDR dataset allows methods to be evaluated by comparing predictions to human judgements ("lighter", "same as", "darker"). The best modern intrinsic image methods learn a map from image to albedo using rendered models and human judgements. This is convenient for practical methods, but cannot explain how a visual agent without geometric, surface and illumination models and a renderer could learn to recover intrinsic images. This paper describes a method that learns intrinsic image decomposition without seeing WHDR annotations, rendered data, or ground truth data. The method relies on paradigms - fake albedos and fake shading fields - together with a novel smoothing procedure that ensures good behavior at short scales on real images. Long scale error is controlled by averaging. Our method achieves WHDR scores competitive with those of strong recent methods allowed to see training WHDR annotations, rendered data, and ground truth data. Because our method is unsupervised, we can compute estimates of the test/train variance of WHDR scores; these are quite large, and it is unsafe to rely small differences in reported WHDR.