Convex Decomposition of Indoor Scenes

TL;DR

This paper presents a method for decomposing complex indoor scenes into a small set of convex primitives using learned regression and refinement techniques, enabling efficient scene abstraction and analysis.

Contribution

The authors introduce a novel convex decomposition approach for indoor scenes that combines learned parsing with refinement, improving scene representation efficiency.

Findings

Primitive representation error comparable to single-image depth prediction

Effective parsing of cluttered indoor scenes into convex primitives

Method improves scene abstraction and analysis accuracy

Abstract

We describe a method to parse a complex, cluttered indoor scene into primitives which offer a parsimonious abstraction of scene structure. Our primitives are simple convexes. Our method uses a learned regression procedure to parse a scene into a fixed number of convexes from RGBD input, and can optionally accept segmentations to improve the decomposition. The result is then polished with a descent method which adjusts the convexes to produce a very good fit, and greedily removes superfluous primitives. Because the entire scene is parsed, we can evaluate using traditional depth, normal, and segmentation error metrics. Our evaluation procedure demonstrates that the error from our primitive representation is comparable to that of predicting depth from a single image.