CVJul 22, 2024

TreeSBA: Tree-Transformer for Self-Supervised Sequential Brick Assembly

arXiv:2407.15648v12 citationsh-index: 14
Originality Incremental advance
AI Analysis

This work addresses the challenge of costly action labels in sequential brick assembly for robotics or design applications, offering an incremental improvement over existing methods.

The paper tackles the problem of inferring step-wise actions for 3D brick assembly from images by proposing a class-agnostic tree-transformer framework that uses synthetic-to-real transfer learning with self-supervision, achieving improvements of 7.8% and 11.3% in mIoU on MNIST and ModelNet Construction datasets without real data annotations.

Inferring step-wise actions to assemble 3D objects with primitive bricks from images is a challenging task due to complex constraints and the vast number of possible combinations. Recent studies have demonstrated promising results on sequential LEGO brick assembly through the utilization of LEGO-Graph modeling to predict sequential actions. However, existing approaches are class-specific and require significant computational and 3D annotation resources. In this work, we first propose a computationally efficient breadth-first search (BFS) LEGO-Tree structure to model the sequential assembly actions by considering connections between consecutive layers. Based on the LEGO-Tree structure, we then design a class-agnostic tree-transformer framework to predict the sequential assembly actions from the input multi-view images. A major challenge of the sequential brick assembly task is that the step-wise action labels are costly and tedious to obtain in practice. We mitigate this problem by leveraging synthetic-to-real transfer learning. Specifically, our model is first pre-trained on synthetic data with full supervision from the available action labels. We then circumvent the requirement for action labels in the real data by proposing an action-to-silhouette projection that replaces action labels with input image silhouettes for self-supervision. Without any annotation on the real data, our model outperforms existing methods with 3D supervision by 7.8% and 11.3% in mIoU on the MNIST and ModelNet Construction datasets, respectively.

Code Implementations1 repo
Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes