MetaSDF: Meta-learning Signed Distance Functions

TL;DR

MetaSDF introduces a meta-learning approach for neural implicit shape representations, enabling faster test-time inference and improved generalization over existing encoder-decoder methods.

Contribution

It formalizes shape space learning as a meta-learning problem and applies gradient-based meta-learning algorithms, achieving faster inference and better performance.

Findings

MetaSDF matches auto-decoder performance.

It is an order of magnitude faster at test time.

Outperforms pooling-based encoder-decoder methods.

Abstract

Neural implicit shape representations are an emerging paradigm that offers many potential benefits over conventional discrete representations, including memory efficiency at a high spatial resolution. Generalizing across shapes with such neural implicit representations amounts to learning priors over the respective function space and enables geometry reconstruction from partial or noisy observations. Existing generalization methods rely on conditioning a neural network on a low-dimensional latent code that is either regressed by an encoder or jointly optimized in the auto-decoder framework. Here, we formalize learning of a shape space as a meta-learning problem and leverage gradient-based meta-learning algorithms to solve this task. We demonstrate that this approach performs on par with auto-decoder based approaches while being an order of magnitude faster at test-time inference. We further demonstrate that the proposed gradient-based method outperforms encoder-decoder based methods that leverage pooling-based set encoders.