Training Data Generating Networks: Shape Reconstruction via Bi-level Optimization

TL;DR

This paper introduces a novel 3D shape reconstruction method that uses a training data generating network within a bi-level optimization framework, linking 3D shape analysis with few-shot learning.

Contribution

It proposes a new 3D shape representation approach that employs bi-level optimization to jointly train data generating networks for improved shape reconstruction.

Findings

Outperforms recent methods on standard 3D shape reconstruction benchmarks.

Establishes a connection between 3D shape analysis and few-shot learning.

Demonstrates the effectiveness of bi-level optimization in training data generation for shape reconstruction.

Abstract

We propose a novel 3d shape representation for 3d shape reconstruction from a single image. Rather than predicting a shape directly, we train a network to generate a training set which will be fed into another learning algorithm to define the shape. The nested optimization problem can be modeled by bi-level optimization. Specifically, the algorithms for bi-level optimization are also being used in meta learning approaches for few-shot learning. Our framework establishes a link between 3D shape analysis and few-shot learning. We combine training data generating networks with bi-level optimization algorithms to obtain a complete framework for which all components can be jointly trained. We improve upon recent work on standard benchmarks for 3d shape reconstruction.