Learning to Infer and Execute 3D Shape Programs

TL;DR

This paper introduces a neural approach to infer and execute 3D shape programs, capturing both low-level geometry and high-level structural priors, enabling more accurate and plausible 3D shape reconstructions from raw data and images.

Contribution

It presents a novel neural framework that learns to infer and execute 3D shape programs from unannotated shapes, integrating bottom-up recognition with top-down symbolic structures.

Findings

Accurately infers complex 3D shape programs from raw shapes

Reconstructs shapes more accurately and plausibly from images

Self-supervised learning enables training without explicit annotations

Abstract

Human perception of 3D shapes goes beyond reconstructing them as a set of points or a composition of geometric primitives: we also effortlessly understand higher-level shape structure such as the repetition and reflective symmetry of object parts. In contrast, recent advances in 3D shape sensing focus more on low-level geometry but less on these higher-level relationships. In this paper, we propose 3D shape programs, integrating bottom-up recognition systems with top-down, symbolic program structure to capture both low-level geometry and high-level structural priors for 3D shapes. Because there are no annotations of shape programs for real shapes, we develop neural modules that not only learn to infer 3D shape programs from raw, unannotated shapes, but also to execute these programs for shape reconstruction. After initial bootstrapping, our end-to-end differentiable model learns 3D shape programs by reconstructing shapes in a self-supervised manner. Experiments demonstrate that our model accurately infers and executes 3D shape programs for highly complex shapes from various categories. It can also be integrated with an image-to-shape module to infer 3D shape programs directly from an RGB image, leading to 3D shape reconstructions that are both more accurate and more physically plausible.