# Superquadrics Revisited: Learning 3D Shape Parsing beyond Cuboids

## TL;DR

This paper introduces a learning-based method for 3D shape parsing that uses superquadrics instead of cuboids, offering more expressive and detailed representations with an analytical Chamfer loss solution.

## Contribution

It presents a novel approach that replaces cuboids with superquadrics for 3D shape parsing, improving expressiveness and learning efficiency without supervision.

## Key findings

- More expressive 3D scene parses achieved
- Easier learning process than cuboid-based methods
- Effective on ShapeNet and SURREAL datasets

## Abstract

Abstracting complex 3D shapes with parsimonious part-based representations has been a long standing goal in computer vision. This paper presents a learning-based solution to this problem which goes beyond the traditional 3D cuboid representation by exploiting superquadrics as atomic elements. We demonstrate that superquadrics lead to more expressive 3D scene parses while being easier to learn than 3D cuboid representations. Moreover, we provide an analytical solution to the Chamfer loss which avoids the need for computational expensive reinforcement learning or iterative prediction. Our model learns to parse 3D objects into consistent superquadric representations without supervision. Results on various ShapeNet categories as well as the SURREAL human body dataset demonstrate the flexibility of our model in capturing fine details and complex poses that could not have been modelled using cuboids.