Manual-PA: Learning 3D Part Assembly from Instruction Diagrams

TL;DR

Manual-PA leverages instruction diagrams and transformer models to improve 3D furniture part assembly by predicting assembly order and part poses, showing significant performance gains and real-world generalization.

Contribution

The paper introduces a novel transformer-based framework that uses diagram cues to split assembly into discrete and continuous phases, enhancing learning efficiency.

Findings

Significant improvement over state-of-the-art methods.

Strong generalization to real-world IKEA furniture assembly.

Effective use of diagram cues for assembly prediction.

Abstract

Assembling furniture amounts to solving the discrete-continuous optimization task of selecting the furniture parts to assemble and estimating their connecting poses in a physically realistic manner. The problem is hampered by its combinatorially large yet sparse solution space thus making learning to assemble a challenging task for current machine learning models. In this paper, we attempt to solve this task by leveraging the assembly instructions provided in diagrammatic manuals that typically accompany the furniture parts. Our key insight is to use the cues in these diagrams to split the problem into discrete and continuous phases. Specifically, we present Manual-PA, a transformer-based instruction Manual-guided 3D Part Assembly framework that learns to semantically align 3D parts with their illustrations in the manuals using a contrastive learning backbone towards predicting the assembly order and infers the 6D pose of each part via relating it to the final furniture depicted in the manual. To validate the efficacy of our method, we conduct experiments on the benchmark PartNet dataset. Our results show that using the diagrams and the order of the parts lead to significant improvements in assembly performance against the state of the art. Further, Manual-PA demonstrates strong generalization to real-world IKEA furniture assembly on the IKEA-Manual dataset.