3DIAS: 3D Shape Reconstruction with Implicit Algebraic Surfaces

TL;DR

This paper introduces a novel implicit algebraic surface primitive for 3D shape reconstruction, leveraging deep learning to produce complex, high-resolution shapes from single RGB images, outperforming existing methods.

Contribution

The paper proposes a constrained implicit algebraic surface primitive with few learnable coefficients and a neural network to generate these primitives, enhancing shape resolution and representation power.

Findings

Outperforms state-of-the-art in 3D shape reconstruction from RGB images

Enables unsupervised semantic segmentation of 3D shapes

Demonstrates high geometrical complexity with fewer primitives

Abstract

3D Shape representation has substantial effects on 3D shape reconstruction. Primitive-based representations approximate a 3D shape mainly by a set of simple implicit primitives, but the low geometrical complexity of the primitives limits the shape resolution. Moreover, setting a sufficient number of primitives for an arbitrary shape is challenging. To overcome these issues, we propose a constrained implicit algebraic surface as the primitive with few learnable coefficients and higher geometrical complexities and a deep neural network to produce these primitives. Our experiments demonstrate the superiorities of our method in terms of representation power compared to the state-of-the-art methods in single RGB image 3D shape reconstruction. Furthermore, we show that our method can semantically learn segments of 3D shapes in an unsupervised manner. The code is publicly available from https://myavartanoo.github.io/3dias/ .