De-rendering 3D Objects in the Wild

TL;DR

This paper introduces a weakly supervised method for decomposing 2D images into detailed 3D representations, including shape, material, and lighting, using minimal initial shape supervision, suitable for diverse real-world objects.

Contribution

It presents a novel weakly supervised approach that leverages rough shape estimates for 3D decomposition from single images, enabling generalization to unseen categories.

Findings

Successfully de-decomposes images into 3D shape, material, and lighting.

Generalizes well to unseen object categories.

Provides a synthetic dataset for quantitative evaluation.

Abstract

With increasing focus on augmented and virtual reality applications (XR) comes the demand for algorithms that can lift objects from images and videos into representations that are suitable for a wide variety of related 3D tasks. Large-scale deployment of XR devices and applications means that we cannot solely rely on supervised learning, as collecting and annotating data for the unlimited variety of objects in the real world is infeasible. We present a weakly supervised method that is able to decompose a single image of an object into shape (depth and normals), material (albedo, reflectivity and shininess) and global lighting parameters. For training, the method only relies on a rough initial shape estimate of the training objects to bootstrap the learning process. This shape supervision can come for example from a pretrained depth network or - more generically - from a traditional structure-from-motion pipeline. In our experiments, we show that the method can successfully de-render 2D images into a decomposed 3D representation and generalizes to unseen object categories. Since in-the-wild evaluation is difficult due to the lack of ground truth data, we also introduce a photo-realistic synthetic test set that allows for quantitative evaluation.