Bayesian Meta-Learning for Few-Shot 3D Shape Completion

Masanori Koyama, Toshiki Nakanishi, Shin-ichi Maeda, Vitor Campagnolo Guizilini, Adrien Gaidon

Keywords: shape completion, Meta-learning, Few-shot, 3D reconstruction

Abstract: Estimating the 3D shape of real-world objects is a key perceptual challenge. It requires going from partial observations, which are often too sparse and incomprehensible for the human eye, to detailed shape representations that vary significantly across categories and instances. We propose to cast shape completion as a Bayesian meta-learning problem to facilitate the transfer of knowledge learned from observing one object into estimating the shape of another object. To combine the Bayesian framework with an approach that uses implicit 3D object representation, we introduce an encoder that describes the posterior distribution of a latent representation conditioned on sparse point clouds. With its ability to isolate object-specific properties from object-agnostic properties, our meta-learning algorithm enables accurate shape completion of newly-encountered objects from sparse observations. We demonstrate the efficacy of our proposed method with experimental results on the standard ShapeNet and ICL-NUIM benchmarks.

Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics

One-sentence Summary: We propose a novel shape completion algorithm that uses Bayesian meta-learning to improve generalization and performance on sparse datasets.

Reviewed Version (pdf): https://openreview.net/references/pdf?id=viZZKDLuGV